1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
3362
3363
3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
3390
3391
3392
3393
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3489
3490
3491
3492
3493
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
3528
3529
3530
3531
3532
3533
3534
3535
3536
3537
3538
3539
3540
3541
3542
3543
3544
3545
3546
3547
3548
3549
3550
3551
3552
3553
3554
3555
3556
3557
3558
3559
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
3571
3572
3573
3574
3575
3576
3577
3578
3579
3580
3581
3582
3583
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
3594
3595
3596
3597
3598
3599
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
3636
3637
3638
3639
3640
3641
3642
3643
3644
3645
3646
3647
3648
3649
3650
3651
3652
3653
3654
3655
3656
3657
3658
3659
3660
3661
3662
3663
3664
3665
3666
3667
3668
3669
3670
3671
3672
3673
3674
3675
3676
3677
3678
3679
3680
3681
3682
3683
3684
3685
3686
3687
3688
3689
3690
3691
3692
3693
3694
3695
3696
3697
3698
3699
3700
3701
3702
3703
3704
3705
3706
3707
3708
3709
3710
3711
3712
3713
3714
3715
3716
3717
3718
3719
3720
3721
3722
3723
3724
3725
3726
3727
3728
3729
3730
3731
3732
3733
3734
3735
3736
3737
3738
3739
3740
3741
3742
3743
3744
3745
3746
3747
3748
3749
3750
3751
3752
3753
3754
3755
3756
3757
3758
3759
3760
3761
3762
3763
3764
3765
3766
3767
3768
3769
3770
3771
3772
3773
3774
3775
3776
3777
3778
3779
3780
3781
3782
3783
3784
3785
3786
3787
3788
3789
3790
3791
3792
3793
3794
3795
3796
3797
3798
3799
3800
3801
3802
3803
3804
3805
3806
3807
3808
3809
3810
3811
3812
3813
3814
3815
3816
3817
3818
3819
3820
3821
3822
3823
3824
3825
3826
3827
3828
3829
3830
3831
3832
3833
3834
3835
3836
3837
3838
3839
3840
3841
3842
3843
3844
3845
3846
3847
3848
3849
3850
3851
3852
3853
3854
3855
3856
3857
3858
3859
3860
3861
3862
3863
3864
3865
3866
3867
3868
3869
3870
3871
3872
3873
3874
3875
3876
3877
3878
3879
3880
3881
3882
3883
3884
3885
3886
3887
3888
3889
3890
3891
3892
3893
3894
3895
3896
3897
3898
3899
3900
3901
3902
3903
3904
3905
3906
3907
3908
3909
3910
3911
3912
3913
3914
3915
3916
3917
3918
3919
3920
3921
3922
3923
3924
3925
3926
3927
3928
3929
3930
3931
3932
3933
3934
3935
3936
3937
3938
3939
3940
3941
3942
3943
3944
3945
3946
3947
3948
3949
3950
3951
3952
3953
3954
3955
3956
3957
3958
3959
3960
3961
3962
3963
3964
3965
3966
3967
3968
3969
3970
3971
3972
3973
3974
3975
3976
3977
3978
3979
3980
3981
3982
3983
3984
3985
3986
3987
3988
3989
3990
3991
3992
3993
3994
3995
3996
3997
3998
3999
4000
4001
4002
4003
4004
4005
4006
4007
4008
4009
4010
4011
4012
4013
4014
4015
4016
4017
4018
4019
4020
4021
4022
4023
4024
4025
4026
4027
4028
4029
4030
4031
4032
4033
4034
4035
4036
4037
4038
4039
4040
4041
4042
4043
4044
4045
4046
4047
4048
4049
4050
4051
4052
4053
4054
4055
4056
4057
4058
4059
4060
4061
4062
4063
4064
4065
4066
4067
4068
4069
4070
4071
4072
4073
4074
4075
4076
4077
4078
4079
4080
4081
4082
4083
4084
4085
4086
4087
4088
4089
4090
4091
4092
4093
4094
4095
4096
4097
4098
4099
4100
4101
4102
4103
4104
4105
4106
4107
4108
4109
4110
4111
4112
4113
4114
4115
4116
4117
4118
4119
4120
4121
4122
4123
4124
4125
4126
4127
4128
4129
4130
4131
4132
4133
4134
4135
4136
4137
4138
4139
4140
4141
4142
4143
4144
4145
4146
4147
4148
4149
4150
4151
4152
4153
4154
4155
4156
4157
4158
4159
4160
4161
4162
4163
4164
4165
4166
4167
4168
4169
4170
4171
4172
4173
4174
4175
4176
4177
4178
4179
4180
4181
4182
4183
4184
4185
4186
4187
4188
4189
4190
4191
4192
4193
4194
4195
4196
4197
4198
4199
4200
4201
4202
4203
4204
4205
4206
4207
4208
4209
4210
4211
4212
4213
4214
4215
4216
4217
4218
4219
4220
4221
4222
4223
4224
4225
4226
4227
4228
4229
4230
4231
4232
4233
4234
4235
4236
4237
4238
4239
4240
4241
4242
4243
4244
4245
4246
4247
4248
4249
4250
4251
4252
4253
4254
4255
4256
4257
4258
4259
4260
4261
4262
4263
4264
4265
4266
4267
4268
4269
4270
4271
4272
4273
4274
4275
4276
4277
4278
4279
4280
4281
4282
4283
4284
4285
4286
4287
4288
4289
4290
4291
4292
4293
4294
4295
4296
4297
4298
4299
4300
4301
4302
4303
4304
4305
4306
4307
4308
4309
4310
4311
4312
4313
4314
4315
4316
4317
4318
4319
4320
4321
4322
4323
4324
4325
4326
4327
4328
4329
4330
4331
4332
4333
4334
4335
4336
4337
4338
4339
4340
4341
4342
4343
4344
4345
4346
4347
4348
4349
4350
4351
4352
4353
4354
4355
4356
4357
4358
4359
4360
4361
4362
4363
4364
4365
4366
4367
4368
4369
4370
4371
4372
4373
4374
4375
4376
4377
4378
4379
4380
4381
4382
4383
4384
4385
4386
4387
4388
4389
4390
4391
4392
4393
4394
4395
4396
4397
4398
4399
4400
4401
4402
4403
4404
4405
4406
4407
4408
4409
4410
4411
4412
4413
4414
4415
4416
4417
4418
4419
4420
4421
4422
4423
4424
4425
4426
4427
4428
4429
4430
4431
4432
4433
4434
4435
4436
4437
4438
4439
4440
4441
4442
4443
4444
4445
4446
4447
4448
4449
4450
4451
4452
4453
4454
4455
4456
4457
4458
4459
4460
4461
4462
4463
4464
4465
4466
4467
4468
4469
4470
4471
4472
4473
4474
4475
4476
4477
4478
4479
4480
4481
4482
4483
4484
4485
4486
4487
4488
4489
4490
4491
4492
4493
4494
4495
4496
4497
4498
4499
4500
4501
4502
4503
4504
4505
4506
4507
4508
4509
4510
4511
4512
4513
4514
4515
4516
4517
4518
4519
4520
4521
4522
4523
4524
4525
4526
4527
4528
4529
4530
4531
4532
4533
4534
4535
4536
4537
4538
4539
4540
4541
4542
4543
4544
4545
4546
4547
4548
4549
4550
4551
4552
4553
4554
4555
4556
4557
4558
4559
4560
4561
4562
4563
4564
4565
4566
4567
4568
4569
4570
4571
4572
4573
4574
4575
4576
4577
4578
4579
4580
4581
4582
4583
4584
4585
4586
4587
4588
4589
4590
4591
4592
4593
4594
4595
4596
4597
4598
4599
4600
4601
4602
4603
4604
4605
4606
4607
4608
4609
4610
4611
4612
4613
4614
4615
4616
4617
4618
4619
4620
4621
4622
4623
4624
4625
4626
4627
4628
4629
4630
4631
4632
4633
4634
4635
4636
4637
4638
4639
4640
4641
4642
4643
4644
4645
4646
4647
4648
4649
4650
4651
4652
4653
4654
4655
4656
4657
4658
4659
4660
4661
4662
4663
4664
4665
4666
4667
4668
4669
4670
4671
4672
4673
4674
4675
4676
4677
4678
4679
4680
4681
4682
4683
4684
4685
4686
4687
4688
4689
4690
4691
4692
4693
4694
4695
4696
4697
4698
4699
4700
4701
4702
4703
4704
4705
4706
4707
4708
4709
4710
4711
4712
4713
4714
4715
4716
4717
4718
4719
4720
4721
4722
4723
4724
4725
4726
4727
4728
4729
4730
4731
4732
4733
4734
4735
4736
4737
4738
4739
4740
4741
4742
4743
4744
4745
4746
4747
4748
4749
4750
4751
4752
4753
4754
4755
4756
4757
4758
4759
4760
4761
4762
4763
4764
4765
4766
4767
4768
4769
4770
4771
4772
4773
4774
4775
4776
4777
4778
4779
4780
4781
4782
4783
4784
4785
4786
4787
4788
4789
4790
4791
4792
4793
4794
4795
4796
4797
4798
4799
4800
4801
4802
4803
4804
4805
4806
4807
4808
4809
4810
4811
4812
4813
4814
4815
4816
4817
4818
4819
4820
4821
4822
4823
4824
4825
4826
4827
4828
4829
4830
4831
4832
4833
4834
4835
4836
4837
4838
4839
4840
4841
4842
4843
4844
4845
4846
4847
4848
4849
4850
4851
4852
4853
4854
4855
4856
4857
4858
4859
4860
4861
4862
4863
4864
4865
4866
4867
4868
4869
4870
4871
4872
4873
4874
4875
4876
4877
4878
4879
4880
4881
4882
4883
4884
4885
4886
4887
4888
4889
4890
4891
4892
4893
4894
4895
4896
4897
4898
4899
4900
4901
4902
4903
4904
4905
4906
4907
4908
4909
4910
4911
4912
4913
4914
4915
4916
4917
4918
4919
4920
4921
4922
4923
4924
4925
4926
4927
4928
4929
4930
4931
4932
4933
4934
4935
4936
4937
4938
4939
4940
4941
4942
4943
4944
4945
4946
4947
4948
4949
4950
4951
4952
4953
4954
4955
4956
4957
4958
4959
4960
4961
4962
4963
4964
4965
4966
4967
4968
4969
4970
4971
4972
4973
4974
4975
4976
4977
4978
4979
4980
4981
4982
4983
4984
4985
4986
4987
4988
4989
4990
4991
4992
4993
4994
4995
4996
4997
4998
4999
5000
5001
5002
5003
5004
5005
5006
5007
5008
5009
5010
5011
5012
5013
5014
5015
5016
5017
5018
5019
5020
5021
5022
5023
5024
5025
5026
5027
5028
5029
5030
5031
5032
5033
5034
5035
5036
5037
5038
5039
5040
5041
5042
5043
5044
5045
5046
5047
5048
5049
5050
5051
5052
5053
5054
5055
5056
5057
5058
5059
5060
5061
5062
5063
5064
5065
5066
5067
5068
5069
5070
5071
5072
5073
5074
5075
5076
5077
5078
5079
5080
5081
5082
5083
5084
5085
5086
5087
5088
5089
5090
5091
5092
5093
5094
5095
5096
5097
5098
5099
5100
5101
5102
5103
5104
5105
5106
5107
5108
5109
5110
5111
5112
5113
5114
5115
5116
5117
5118
5119
5120
5121
5122
5123
5124
5125
5126
5127
5128
5129
5130
5131
5132
5133
5134
5135
5136
5137
5138
5139
5140
5141
5142
5143
5144
5145
5146
5147
5148
5149
5150
5151
5152
5153
5154
5155
5156
5157
5158
5159
5160
5161
5162
5163
5164
5165
5166
5167
5168
5169
5170
5171
5172
5173
5174
5175
5176
5177
5178
5179
5180
5181
5182
5183
5184
5185
5186
5187
5188
5189
5190
5191
5192
5193
5194
5195
5196
5197
5198
5199
5200
5201
5202
5203
5204
5205
5206
5207
5208
5209
5210
5211
5212
5213
5214
5215
5216
5217
5218
5219
5220
5221
5222
5223
5224
5225
5226
5227
5228
5229
5230
5231
5232
5233
5234
5235
5236
5237
5238
5239
5240
5241
5242
5243
5244
5245
5246
5247
5248
5249
5250
5251
5252
5253
5254
5255
5256
5257
5258
5259
5260
5261
5262
5263
5264
5265
5266
5267
5268
5269
5270
5271
5272
5273
5274
5275
5276
5277
5278
5279
5280
5281
5282
5283
5284
5285
5286
5287
5288
5289
5290
5291
5292
5293
5294
5295
5296
5297
5298
5299
5300
5301
5302
5303
5304
5305
5306
5307
5308
5309
5310
5311
5312
5313
5314
5315
5316
5317
5318
5319
5320
5321
5322
5323
5324
5325
5326
5327
5328
5329
5330
5331
5332
5333
5334
5335
5336
5337
5338
5339
5340
5341
5342
5343
5344
5345
5346
5347
5348
5349
5350
5351
5352
5353
5354
5355
5356
5357
5358
5359
5360
5361
5362
5363
5364
5365
5366
5367
5368
5369
5370
5371
5372
5373
5374
5375
5376
5377
5378
5379
5380
5381
5382
5383
5384
5385
5386
5387
5388
5389
5390
5391
5392
5393
5394
5395
5396
5397
5398
5399
5400
5401
5402
5403
5404
5405
5406
5407
5408
5409
5410
5411
5412
5413
5414
5415
5416
5417
5418
5419
5420
5421
5422
5423
5424
5425
5426
5427
5428
5429
5430
5431
5432
5433
5434
5435
5436
5437
5438
5439
5440
5441
5442
5443
5444
5445
5446
5447
5448
5449
5450
5451
5452
5453
5454
5455
5456
5457
5458
5459
5460
5461
5462
5463
5464
5465
5466
5467
5468
5469
5470
5471
5472
5473
5474
5475
5476
5477
5478
5479
5480
5481
5482
5483
5484
5485
5486
5487
5488
5489
5490
5491
5492
5493
5494
5495
5496
5497
5498
5499
5500
5501
5502
5503
5504
5505
5506
5507
5508
5509
5510
5511
5512
5513
5514
5515
5516
5517
5518
5519
5520
5521
5522
5523
5524
5525
5526
5527
5528
5529
5530
5531
5532
5533
5534
5535
5536
5537
5538
5539
5540
5541
5542
5543
5544
5545
5546
5547
5548
5549
5550
5551
5552
5553
5554
5555
5556
5557
5558
5559
5560
5561
5562
5563
5564
5565
5566
5567
5568
5569
5570
5571
5572
5573
5574
5575
5576
5577
5578
5579
5580
5581
5582
5583
5584
5585
5586
5587
5588
5589
5590
5591
5592
5593
5594
5595
5596
5597
5598
5599
5600
5601
5602
5603
5604
5605
5606
5607
5608
5609
5610
5611
5612
5613
5614
5615
5616
5617
5618
5619
5620
5621
5622
5623
5624
5625
5626
5627
5628
5629
5630
5631
5632
5633
5634
5635
5636
5637
5638
5639
5640
5641
5642
5643
5644
5645
5646
5647
5648
5649
5650
5651
5652
5653
5654
5655
5656
5657
5658
5659
5660
5661
5662
5663
5664
5665
5666
5667
5668
5669
5670
5671
5672
5673
5674
5675
5676
5677
5678
5679
5680
5681
5682
5683
5684
5685
5686
5687
5688
5689
5690
5691
5692
5693
5694
5695
5696
5697
5698
5699
5700
5701
5702
5703
5704
5705
5706
5707
5708
5709
5710
5711
5712
5713
5714
5715
5716
5717
5718
5719
5720
5721
5722
5723
5724
5725
5726
5727
5728
5729
5730
5731
5732
5733
5734
5735
5736
5737
5738
5739
5740
5741
5742
5743
5744
5745
5746
5747
5748
5749
5750
5751
5752
5753
5754
5755
5756
5757
5758
5759
5760
5761
5762
5763
5764
5765
5766
5767
5768
5769
5770
5771
5772
5773
5774
5775
5776
5777
5778
5779
5780
5781
5782
5783
5784
5785
5786
5787
5788
5789
5790
5791
5792
5793
5794
5795
5796
5797
5798
5799
5800
5801
5802
5803
5804
5805
5806
5807
5808
5809
5810
5811
5812
5813
5814
5815
5816
5817
5818
5819
5820
5821
5822
5823
5824
5825
5826
5827
5828
5829
5830
5831
5832
5833
5834
5835
5836
5837
5838
5839
5840
5841
5842
5843
5844
5845
5846
5847
5848
5849
5850
5851
5852
5853
5854
5855
5856
5857
5858
5859
5860
5861
5862
5863
5864
5865
5866
5867
5868
5869
5870
5871
5872
5873
5874
5875
5876
5877
5878
5879
5880
5881
5882
5883
5884
5885
5886
5887
5888
5889
5890
5891
5892
5893
5894
5895
5896
5897
5898
5899
5900
5901
5902
5903
5904
5905
5906
5907
5908
5909
5910
5911
5912
5913
5914
5915
5916
5917
5918
5919
5920
5921
5922
5923
5924
5925
5926
5927
5928
5929
5930
5931
5932
5933
5934
5935
5936
5937
5938
5939
5940
5941
5942
5943
5944
5945
5946
5947
5948
5949
5950
5951
5952
5953
5954
5955
5956
5957
5958
5959
5960
5961
5962
5963
5964
5965
5966
5967
5968
5969
5970
5971
5972
5973
5974
5975
5976
5977
5978
5979
5980
5981
5982
5983
5984
5985
5986
5987
5988
5989
5990
5991
5992
5993
5994
5995
5996
5997
5998
5999
6000
6001
6002
6003
6004
6005
6006
6007
6008
6009
6010
6011
6012
6013
6014
6015
6016
6017
6018
6019
6020
6021
6022
6023
6024
6025
6026
6027
6028
6029
6030
6031
6032
6033
6034
6035
6036
6037
6038
6039
6040
6041
6042
6043
6044
6045
6046
6047
6048
6049
6050
6051
6052
6053
6054
6055
6056
6057
6058
6059
6060
6061
6062
6063
6064
6065
6066
6067
6068
6069
6070
6071
6072
6073
6074
6075
6076
6077
6078
6079
6080
6081
6082
6083
6084
6085
6086
6087
6088
6089
6090
6091
6092
6093
6094
6095
6096
6097
6098
6099
6100
6101
6102
6103
6104
6105
6106
6107
6108
6109
6110
6111
6112
6113
6114
6115
6116
6117
6118
6119
6120
6121
6122
6123
6124
6125
6126
6127
6128
6129
6130
6131
6132
6133
6134
6135
6136
6137
6138
6139
6140
6141
6142
6143
6144
6145
6146
6147
6148
6149
6150
6151
6152
6153
6154
6155
6156
6157
6158
6159
6160
6161
6162
6163
6164
6165
6166
6167
6168
6169
6170
6171
6172
6173
6174
6175
6176
6177
6178
6179
6180
6181
6182
6183
6184
6185
6186
6187
6188
6189
6190
6191
6192
6193
6194
6195
6196
6197
6198
6199
6200
6201
6202
6203
6204
6205
6206
6207
6208
6209
6210
6211
6212
6213
6214
6215
6216
6217
6218
6219
6220
6221
6222
6223
6224
6225
6226
6227
6228
6229
6230
6231
6232
6233
6234
6235
6236
6237
6238
6239
6240
6241
6242
6243
6244
6245
6246
6247
6248
6249
6250
6251
6252
6253
6254
6255
6256
6257
6258
6259
6260
6261
6262
6263
6264
6265
6266
6267
6268
6269
6270
6271
6272
6273
6274
6275
6276
6277
6278
6279
6280
6281
6282
6283
6284
6285
6286
6287
6288
6289
6290
6291
6292
6293
6294
6295
6296
6297
6298
6299
6300
6301
6302
6303
6304
6305
6306
6307
6308
6309
6310
6311
6312
6313
6314
6315
6316
6317
6318
6319
6320
6321
6322
6323
6324
6325
6326
6327
6328
6329
6330
6331
6332
6333
6334
6335
6336
6337
6338
6339
6340
6341
6342
6343
6344
6345
6346
6347
6348
6349
6350
6351
6352
6353
6354
6355
6356
6357
6358
6359
6360
6361
6362
6363
6364
6365
6366
6367
6368
6369
6370
6371
6372
6373
6374
6375
6376
6377
6378
6379
6380
6381
6382
6383
6384
6385
6386
6387
6388
6389
6390
6391
6392
6393
6394
6395
6396
6397
6398
6399
6400
6401
6402
6403
6404
6405
6406
6407
6408
6409
6410
6411
6412
6413
6414
6415
6416
6417
6418
6419
6420
6421
6422
6423
6424
6425
6426
6427
6428
6429
6430
6431
6432
6433
6434
6435
6436
6437
6438
6439
6440
6441
6442
6443
6444
6445
6446
6447
6448
6449
6450
6451
6452
6453
6454
6455
6456
6457
6458
6459
6460
6461
6462
6463
6464
6465
6466
6467
6468
6469
6470
6471
6472
6473
6474
6475
6476
6477
6478
6479
6480
6481
6482
6483
6484
6485
6486
6487
6488
6489
6490
6491
6492
6493
6494
6495
6496
6497
6498
6499
6500
6501
6502
6503
6504
6505
6506
6507
6508
6509
6510
6511
6512
6513
6514
6515
6516
6517
6518
6519
6520
6521
6522
6523
6524
6525
6526
6527
6528
6529
6530
6531
6532
6533
6534
6535
6536
6537
6538
6539
6540
6541
6542
6543
6544
6545
6546
6547
6548
6549
6550
6551
6552
6553
6554
6555
6556
6557
6558
6559
6560
6561
6562
6563
6564
6565
6566
6567
6568
6569
6570
6571
6572
6573
6574
6575
6576
6577
6578
6579
6580
6581
6582
6583
6584
6585
6586
6587
6588
6589
6590
6591
6592
6593
6594
6595
6596
6597
6598
6599
6600
6601
6602
6603
6604
6605
6606
6607
6608
6609
6610
6611
6612
6613
6614
6615
6616
6617
6618
6619
6620
6621
6622
6623
6624
6625
6626
6627
6628
6629
6630
6631
6632
6633
6634
6635
6636
6637
6638
6639
6640
6641
6642
6643
6644
6645
6646
6647
6648
6649
6650
6651
6652
6653
6654
6655
6656
6657
6658
6659
6660
6661
6662
6663
6664
6665
6666
6667
6668
6669
6670
6671
6672
6673
6674
6675
6676
6677
6678
6679
6680
6681
6682
6683
6684
6685
6686
6687
6688
6689
6690
6691
6692
6693
6694
6695
6696
6697
6698
6699
6700
6701
6702
6703
6704
6705
6706
6707
6708
6709
6710
6711
6712
6713
6714
6715
6716
6717
6718
6719
6720
6721
6722
6723
6724
6725
6726
6727
6728
6729
6730
6731
6732
6733
6734
6735
6736
6737
6738
6739
6740
6741
6742
6743
6744
6745
6746
6747
6748
6749
6750
6751
6752
6753
6754
6755
6756
6757
6758
6759
6760
6761
6762
6763
6764
6765
6766
6767
6768
6769
6770
6771
6772
6773
6774
6775
6776
6777
6778
6779
6780
6781
6782
6783
6784
6785
6786
6787
6788
6789
6790
6791
6792
6793
6794
6795
6796
6797
6798
6799
6800
6801
6802
6803
6804
6805
6806
6807
6808
6809
6810
6811
6812
6813
6814
6815
6816
6817
6818
6819
6820
6821
6822
6823
6824
6825
6826
6827
6828
6829
6830
6831
6832
6833
6834
6835
6836
6837
6838
6839
6840
6841
6842
6843
6844
6845
6846
6847
6848
6849
6850
6851
6852
6853
6854
6855
6856
6857
6858
6859
6860
6861
6862
6863
6864
6865
6866
6867
6868
6869
6870
6871
6872
6873
6874
6875
6876
6877
6878
6879
6880
6881
6882
6883
6884
6885
6886
6887
6888
6889
6890
6891
6892
6893
6894
6895
6896
6897
6898
6899
6900
6901
6902
6903
6904
6905
6906
6907
6908
6909
6910
6911
6912
6913
6914
6915
6916
6917
6918
6919
6920
6921
6922
6923
6924
6925
6926
6927
6928
6929
6930
6931
6932
6933
6934
6935
6936
6937
6938
6939
6940
6941
6942
6943
6944
6945
6946
6947
6948
6949
6950
6951
6952
6953
6954
6955
6956
6957
6958
6959
6960
6961
6962
6963
6964
6965
6966
6967
6968
6969
6970
6971
6972
6973
6974
6975
6976
6977
6978
6979
6980
6981
6982
6983
6984
6985
6986
6987
6988
6989
6990
6991
6992
6993
6994
6995
6996
6997
6998
6999
7000
7001
7002
7003
7004
7005
7006
7007
7008
7009
7010
7011
7012
7013
7014
7015
7016
7017
7018
7019
7020
7021
7022
7023
7024
7025
7026
7027
7028
7029
7030
7031
7032
7033
7034
7035
7036
7037
7038
7039
7040
7041
7042
7043
7044
7045
7046
7047
7048
7049
7050
7051
7052
7053
7054
7055
7056
7057
7058
7059
7060
7061
7062
7063
7064
7065
7066
7067
7068
7069
7070
7071
7072
7073
7074
7075
7076
7077
7078
7079
7080
7081
7082
7083
7084
7085
7086
7087
7088
7089
7090
7091
7092
7093
7094
7095
7096
7097
7098
7099
7100
7101
7102
7103
7104
7105
7106
7107
7108
7109
7110
7111
7112
7113
7114
7115
7116
7117
7118
7119
7120
7121
7122
7123
7124
7125
7126
7127
7128
7129
7130
7131
7132
7133
7134
7135
7136
7137
7138
7139
7140
7141
7142
7143
7144
7145
7146
7147
7148
7149
7150
7151
7152
7153
7154
7155
7156
7157
7158
7159
7160
7161
7162
7163
7164
7165
7166
7167
7168
7169
7170
7171
7172
7173
7174
7175
7176
7177
7178
7179
7180
7181
7182
7183
7184
7185
7186
7187
7188
7189
7190
7191
7192
7193
7194
7195
7196
7197
7198
7199
7200
7201
7202
7203
7204
7205
7206
7207
7208
7209
7210
7211
7212
7213
7214
7215
7216
7217
7218
7219
7220
7221
7222
7223
7224
7225
7226
7227
7228
7229
7230
7231
7232
7233
7234
7235
7236
7237
7238
7239
7240
7241
7242
7243
7244
7245
7246
7247
7248
7249
7250
7251
7252
7253
7254
7255
7256
7257
7258
7259
7260
7261
7262
7263
7264
7265
7266
7267
7268
7269
7270
7271
7272
7273
7274
7275
7276
7277
7278
7279
7280
7281
7282
7283
7284
7285
7286
7287
7288
7289
7290
7291
7292
7293
7294
7295
7296
7297
7298
7299
7300
7301
7302
7303
7304
7305
7306
7307
7308
7309
7310
7311
7312
7313
7314
7315
7316
7317
7318
7319
7320
7321
7322
7323
7324
7325
7326
7327
7328
7329
7330
7331
7332
7333
7334
7335
7336
7337
7338
7339
7340
7341
7342
7343
7344
7345
7346
7347
7348
7349
7350
7351
7352
7353
7354
7355
7356
7357
7358
7359
7360
7361
7362
7363
7364
7365
7366
7367
7368
7369
7370
7371
7372
7373
7374
7375
7376
7377
7378
7379
7380
7381
7382
7383
7384
7385
7386
7387
7388
7389
7390
7391
7392
7393
7394
7395
7396
7397
7398
7399
7400
7401
7402
7403
7404
7405
7406
7407
7408
7409
7410
7411
7412
7413
7414
7415
7416
7417
7418
7419
7420
7421
7422
7423
7424
7425
7426
7427
7428
7429
7430
7431
7432
7433
7434
7435
7436
7437
7438
7439
7440
7441
7442
7443
7444
7445
7446
7447
7448
7449
7450
7451
7452
7453
7454
7455
7456
7457
7458
7459
7460
7461
7462
7463
7464
7465
7466
7467
7468
7469
7470
7471
7472
7473
7474
7475
7476
7477
7478
7479
7480
7481
7482
7483
7484
7485
7486
7487
7488
7489
7490
7491
7492
7493
7494
7495
7496
7497
7498
7499
7500
7501
7502
7503
7504
7505
7506
7507
7508
7509
7510
7511
7512
7513
7514
7515
7516
7517
7518
7519
7520
7521
7522
7523
7524
7525
7526
7527
7528
7529
7530
7531
7532
7533
7534
7535
7536
7537
7538
7539
7540
7541
7542
7543
7544
7545
7546
7547
7548
7549
7550
7551
7552
7553
7554
7555
7556
7557
7558
7559
7560
7561
7562
7563
7564
7565
7566
7567
7568
7569
7570
7571
7572
7573
7574
7575
7576
7577
7578
7579
7580
7581
7582
7583
7584
7585
7586
7587
7588
7589
7590
7591
7592
7593
7594
7595
7596
7597
7598
7599
7600
7601
7602
7603
7604
7605
7606
7607
7608
7609
7610
7611
7612
7613
7614
7615
7616
7617
7618
7619
7620
7621
7622
7623
7624
7625
7626
7627
7628
7629
7630
7631
7632
7633
7634
7635
7636
7637
7638
7639
7640
7641
7642
7643
7644
7645
7646
7647
7648
7649
7650
7651
7652
7653
7654
7655
7656
7657
7658
7659
7660
7661
7662
7663
7664
7665
7666
7667
7668
7669
7670
7671
7672
7673
7674
7675
7676
7677
7678
7679
7680
7681
7682
7683
7684
7685
7686
7687
7688
7689
7690
7691
7692
7693
7694
7695
7696
7697
7698
7699
7700
7701
7702
7703
7704
7705
7706
7707
7708
7709
7710
7711
7712
7713
7714
7715
7716
7717
7718
7719
7720
7721
7722
7723
7724
7725
7726
7727
7728
7729
7730
7731
7732
7733
7734
7735
7736
7737
7738
7739
7740
7741
7742
7743
7744
7745
7746
7747
7748
7749
7750
7751
7752
7753
7754
7755
7756
7757
7758
7759
7760
7761
7762
7763
7764
7765
7766
7767
7768
7769
7770
7771
7772
7773
7774
7775
7776
7777
7778
7779
7780
7781
7782
7783
7784
7785
7786
7787
7788
7789
7790
7791
7792
7793
7794
7795
7796
7797
7798
7799
7800
7801
7802
7803
7804
7805
7806
7807
7808
7809
7810
7811
7812
7813
7814
7815
7816
7817
7818
7819
7820
7821
7822
7823
7824
7825
7826
7827
7828
7829
7830
7831
7832
7833
7834
7835
7836
7837
7838
7839
7840
7841
7842
7843
7844
7845
7846
7847
7848
7849
7850
7851
7852
7853
7854
7855
7856
7857
7858
7859
7860
7861
7862
7863
7864
7865
7866
7867
7868
7869
7870
7871
7872
7873
7874
7875
7876
7877
7878
7879
7880
7881
7882
7883
7884
7885
7886
7887
7888
7889
7890
7891
7892
7893
7894
7895
7896
7897
7898
7899
7900
7901
7902
7903
7904
7905
7906
7907
7908
7909
7910
7911
7912
7913
7914
7915
7916
7917
7918
7919
7920
7921
7922
7923
7924
7925
7926
7927
7928
7929
7930
7931
7932
7933
7934
7935
7936
7937
7938
7939
7940
7941
7942
7943
7944
7945
7946
7947
7948
7949
7950
7951
7952
7953
7954
7955
7956
7957
7958
7959
7960
7961
7962
7963
7964
7965
7966
7967
7968
7969
7970
7971
7972
7973
7974
7975
7976
7977
7978
7979
7980
7981
7982
7983
7984
7985
7986
7987
7988
7989
7990
7991
7992
7993
7994
7995
7996
7997
7998
7999
8000
8001
8002
8003
8004
8005
8006
8007
8008
8009
8010
8011
8012
8013
8014
8015
8016
8017
8018
8019
8020
8021
8022
8023
8024
8025
8026
8027
8028
8029
8030
8031
8032
8033
8034
8035
8036
8037
8038
8039
8040
8041
8042
8043
8044
8045
8046
8047
8048
8049
8050
8051
8052
8053
8054
8055
8056
8057
8058
8059
8060
8061
8062
8063
8064
8065
8066
8067
8068
8069
8070
8071
8072
8073
8074
8075
8076
8077
8078
8079
8080
8081
8082
8083
8084
8085
8086
8087
8088
8089
8090
8091
8092
8093
8094
8095
8096
8097
8098
8099
8100
8101
8102
8103
8104
8105
8106
8107
8108
8109
8110
8111
8112
8113
8114
8115
8116
8117
8118
8119
8120
8121
8122
8123
8124
8125
8126
8127
8128
8129
8130
8131
8132
8133
8134
8135
8136
8137
8138
8139
8140
8141
8142
8143
8144
8145
8146
8147
8148
8149
8150
8151
8152
8153
8154
8155
8156
8157
8158
8159
8160
8161
8162
8163
8164
8165
8166
8167
8168
8169
8170
8171
8172
8173
8174
8175
8176
8177
8178
8179
8180
8181
8182
8183
8184
8185
8186
8187
8188
8189
8190
8191
8192
8193
8194
8195
8196
8197
8198
8199
8200
8201
8202
8203
8204
8205
8206
8207
8208
8209
8210
8211
8212
8213
8214
8215
8216
8217
8218
8219
8220
8221
8222
8223
8224
8225
8226
8227
8228
8229
8230
8231
8232
8233
8234
8235
8236
8237
8238
8239
8240
8241
8242
8243
8244
8245
8246
8247
8248
8249
8250
8251
8252
8253
8254
8255
8256
8257
8258
8259
8260
8261
8262
8263
8264
8265
8266
8267
8268
8269
8270
8271
8272
8273
8274
8275
8276
8277
8278
8279
8280
8281
8282
8283
8284
8285
8286
8287
8288
8289
8290
8291
8292
8293
8294
8295
8296
8297
8298
8299
8300
8301
8302
8303
8304
8305
8306
8307
8308
8309
8310
8311
8312
8313
8314
8315
8316
8317
8318
8319
8320
8321
8322
8323
8324
8325
8326
8327
8328
8329
8330
8331
8332
8333
8334
8335
8336
8337
8338
8339
8340
8341
8342
8343
8344
8345
8346
8347
8348
8349
8350
8351
8352
8353
8354
8355
8356
8357
8358
8359
8360
8361
8362
8363
8364
8365
8366
8367
8368
8369
8370
8371
8372
8373
8374
8375
8376
8377
8378
8379
8380
8381
8382
8383
8384
8385
8386
8387
8388
8389
8390
8391
8392
8393
8394
8395
8396
8397
8398
8399
8400
8401
8402
8403
8404
8405
8406
8407
8408
8409
8410
8411
8412
8413
8414
8415
8416
8417
8418
8419
8420
8421
8422
8423
8424
8425
8426
8427
8428
8429
8430
8431
8432
8433
8434
8435
8436
8437
8438
8439
8440
8441
8442
8443
8444
8445
8446
8447
8448
8449
8450
8451
8452
8453
8454
8455
8456
8457
8458
8459
8460
8461
8462
8463
8464
8465
8466
8467
8468
8469
8470
8471
8472
8473
8474
8475
8476
8477
8478
8479
8480
8481
8482
8483
8484
8485
8486
8487
8488
8489
8490
8491
8492
8493
8494
8495
8496
8497
8498
8499
8500
8501
8502
8503
8504
8505
8506
8507
8508
8509
8510
8511
8512
8513
8514
8515
8516
8517
8518
8519
8520
8521
8522
8523
8524
8525
8526
8527
8528
8529
8530
8531
8532
8533
8534
8535
8536
8537
8538
8539
8540
8541
8542
8543
8544
8545
8546
8547
8548
8549
8550
8551
8552
8553
8554
8555
8556
8557
8558
8559
8560
8561
8562
8563
8564
8565
8566
8567
8568
8569
8570
8571
8572
8573
8574
8575
8576
8577
8578
8579
8580
8581
8582
8583
8584
8585
8586
8587
8588
8589
8590
8591
8592
8593
8594
8595
8596
8597
8598
8599
8600
8601
8602
8603
8604
8605
8606
8607
8608
8609
8610
8611
8612
8613
8614
8615
8616
8617
8618
8619
8620
8621
8622
8623
8624
8625
8626
8627
8628
8629
8630
8631
8632
8633
8634
8635
8636
8637
8638
8639
8640
8641
8642
8643
8644
8645
8646
8647
8648
8649
8650
8651
8652
8653
8654
8655
8656
8657
8658
8659
8660
8661
8662
8663
8664
8665
8666
8667
8668
8669
8670
8671
8672
8673
8674
8675
8676
8677
8678
8679
8680
8681
8682
8683
8684
8685
8686
8687
8688
8689
8690
8691
8692
8693
8694
8695
8696
8697
8698
8699
8700
8701
8702
8703
8704
8705
8706
8707
8708
8709
8710
8711
8712
8713
8714
8715
8716
8717
8718
8719
8720
8721
8722
8723
8724
8725
8726
8727
8728
8729
8730
8731
8732
8733
8734
8735
8736
8737
8738
8739
8740
8741
8742
8743
8744
8745
8746
8747
8748
8749
8750
8751
8752
8753
8754
8755
8756
8757
8758
8759
8760
8761
8762
8763
8764
8765
8766
8767
8768
8769
8770
8771
8772
8773
8774
8775
8776
8777
8778
8779
8780
8781
8782
8783
8784
8785
8786
8787
8788
8789
8790
8791
8792
8793
8794
8795
8796
8797
8798
8799
8800
8801
8802
8803
8804
8805
8806
8807
8808
8809
8810
8811
8812
8813
8814
8815
8816
8817
8818
8819
8820
8821
8822
8823
8824
8825
8826
8827
8828
8829
8830
8831
8832
8833
8834
8835
8836
8837
8838
8839
8840
8841
8842
8843
8844
8845
8846
8847
8848
8849
8850
8851
8852
8853
8854
8855
8856
8857
8858
8859
8860
8861
8862
8863
8864
8865
8866
8867
8868
8869
8870
8871
8872
8873
8874
8875
8876
8877
8878
8879
8880
8881
8882
8883
8884
8885
8886
8887
8888
8889
8890
8891
8892
8893
8894
8895
8896
8897
8898
8899
8900
8901
8902
8903
8904
8905
8906
8907
8908
8909
8910
8911
8912
8913
8914
8915
8916
8917
8918
8919
8920
8921
8922
8923
8924
8925
8926
8927
8928
8929
8930
8931
8932
8933
8934
8935
8936
8937
8938
8939
8940
8941
8942
8943
8944
8945
8946
8947
8948
8949
8950
8951
8952
8953
8954
8955
8956
8957
8958
8959
8960
8961
8962
8963
8964
8965
8966
8967
8968
8969
8970
8971
8972
8973
8974
8975
8976
8977
8978
8979
8980
8981
8982
8983
8984
8985
8986
8987
8988
8989
8990
8991
8992
8993
8994
8995
8996
8997
8998
8999
9000
9001
9002
9003
9004
9005
9006
9007
9008
9009
9010
9011
9012
9013
9014
9015
9016
9017
9018
9019
9020
9021
9022
9023
9024
9025
9026
9027
9028
9029
9030
9031
9032
9033
9034
9035
9036
9037
9038
9039
9040
9041
9042
9043
9044
9045
9046
9047
9048
9049
9050
9051
9052
9053
9054
9055
9056
9057
9058
9059
9060
9061
9062
9063
9064
9065
9066
9067
9068
9069
9070
9071
9072
9073
9074
9075
9076
9077
9078
9079
9080
9081
9082
9083
9084
9085
9086
9087
9088
9089
9090
9091
9092
9093
9094
9095
9096
9097
9098
9099
9100
9101
9102
9103
9104
9105
9106
9107
9108
9109
9110
9111
9112
9113
9114
9115
9116
9117
9118
9119
9120
9121
9122
9123
9124
9125
9126
9127
9128
9129
9130
9131
9132
9133
9134
9135
9136
9137
9138
9139
9140
9141
9142
9143
9144
9145
9146
9147
9148
9149
9150
9151
9152
9153
9154
9155
9156
9157
9158
9159
9160
9161
9162
9163
9164
9165
9166
9167
9168
9169
9170
9171
9172
|
/*
Copyright (c) 2005, 2017, Oracle and/or its affiliates.
Copyright (c) 2009, 2017, MariaDB
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; version 2 of the License.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
/*
This handler was developed by Mikael Ronstrom for version 5.1 of MySQL.
It is an abstraction layer on top of other handlers such as MyISAM,
InnoDB, Federated, Berkeley DB and so forth. Partitioned tables can also
be handled by a storage engine. The current example of this is NDB
Cluster that has internally handled partitioning. This have benefits in
that many loops needed in the partition handler can be avoided.
Partitioning has an inherent feature which in some cases is positive and
in some cases is negative. It splits the data into chunks. This makes
the data more manageable, queries can easily be parallelised towards the
parts and indexes are split such that there are less levels in the
index trees. The inherent disadvantage is that to use a split index
one has to scan all index parts which is ok for large queries but for
small queries it can be a disadvantage.
Partitioning lays the foundation for more manageable databases that are
extremely large. It does also lay the foundation for more parallelism
in the execution of queries. This functionality will grow with later
versions of MySQL.
The partition is setup to use table locks. It implements an partition "SHARE"
that is inserted into a hash by table name. You can use this to store
information of state that any partition handler object will be able to see
if it is using the same table.
Please read the object definition in ha_partition.h before reading the rest
if this file.
*/
#include "mariadb.h"
#include "sql_priv.h"
#include "sql_parse.h" // append_file_to_dir
#include "create_options.h"
#ifdef WITH_PARTITION_STORAGE_ENGINE
#include "ha_partition.h"
#include "sql_table.h" // tablename_to_filename
#include "key.h"
#include "sql_plugin.h"
#include "sql_show.h" // append_identifier
#include "sql_admin.h" // SQL_ADMIN_MSG_TEXT_SIZE
#include "debug_sync.h"
/* First 4 bytes in the .par file is the number of 32-bit words in the file */
#define PAR_WORD_SIZE 4
/* offset to the .par file checksum */
#define PAR_CHECKSUM_OFFSET 4
/* offset to the total number of partitions */
#define PAR_NUM_PARTS_OFFSET 8
/* offset to the engines array */
#define PAR_ENGINES_OFFSET 12
#define PARTITION_ENABLED_TABLE_FLAGS (HA_FILE_BASED | \
HA_REC_NOT_IN_SEQ | \
HA_CAN_REPAIR)
#define PARTITION_DISABLED_TABLE_FLAGS (HA_CAN_GEOMETRY | \
HA_CAN_FULLTEXT | \
HA_DUPLICATE_POS | \
HA_CAN_SQL_HANDLER | \
HA_CAN_INSERT_DELAYED | \
HA_READ_BEFORE_WRITE_REMOVAL |\
HA_CAN_TABLES_WITHOUT_ROLLBACK)
static const char *ha_par_ext= ".par";
/****************************************************************************
MODULE create/delete handler object
****************************************************************************/
static handler *partition_create_handler(handlerton *hton,
TABLE_SHARE *share,
MEM_ROOT *mem_root);
static uint partition_flags();
static uint alter_table_flags(uint flags);
extern "C" int cmp_key_part_id(void *key_p, uchar *ref1, uchar *ref2);
extern "C" int cmp_key_rowid_part_id(void *ptr, uchar *ref1, uchar *ref2);
/*
If frm_error() is called then we will use this to to find out what file
extensions exist for the storage engine. This is also used by the default
rename_table and delete_table method in handler.cc.
*/
static const char *ha_partition_ext[]=
{
ha_par_ext, NullS
};
#ifdef HAVE_PSI_INTERFACE
PSI_mutex_key key_partition_auto_inc_mutex;
static PSI_mutex_info all_partition_mutexes[]=
{
{ &key_partition_auto_inc_mutex, "Partition_share::auto_inc_mutex", 0}
};
static void init_partition_psi_keys(void)
{
const char* category= "partition";
int count;
count= array_elements(all_partition_mutexes);
mysql_mutex_register(category, all_partition_mutexes, count);
}
#endif /* HAVE_PSI_INTERFACE */
static int partition_initialize(void *p)
{
handlerton *partition_hton;
partition_hton= (handlerton *)p;
partition_hton->state= SHOW_OPTION_YES;
partition_hton->db_type= DB_TYPE_PARTITION_DB;
partition_hton->create= partition_create_handler;
partition_hton->partition_flags= partition_flags;
partition_hton->alter_table_flags= alter_table_flags;
partition_hton->flags= HTON_NOT_USER_SELECTABLE |
HTON_HIDDEN |
HTON_TEMPORARY_NOT_SUPPORTED;
partition_hton->tablefile_extensions= ha_partition_ext;
#ifdef HAVE_PSI_INTERFACE
init_partition_psi_keys();
#endif
return 0;
}
/**
Initialize and allocate space for partitions shares.
@param num_parts Number of partitions to allocate storage for.
@return Operation status.
@retval true Failure (out of memory).
@retval false Success.
*/
bool Partition_share::init(uint num_parts)
{
DBUG_ENTER("Partition_share::init");
auto_inc_initialized= false;
partition_name_hash_initialized= false;
next_auto_inc_val= 0;
if (partitions_share_refs.init(num_parts))
{
DBUG_RETURN(true);
}
DBUG_RETURN(false);
}
/*
Create new partition handler
SYNOPSIS
partition_create_handler()
table Table object
RETURN VALUE
New partition object
*/
static handler *partition_create_handler(handlerton *hton,
TABLE_SHARE *share,
MEM_ROOT *mem_root)
{
ha_partition *file= new (mem_root) ha_partition(hton, share);
if (file && file->initialize_partition(mem_root))
{
delete file;
file= 0;
}
return file;
}
/*
HA_CAN_PARTITION:
Used by storage engines that can handle partitioning without this
partition handler
(Partition, NDB)
HA_CAN_UPDATE_PARTITION_KEY:
Set if the handler can update fields that are part of the partition
function.
HA_CAN_PARTITION_UNIQUE:
Set if the handler can handle unique indexes where the fields of the
unique key are not part of the fields of the partition function. Thus
a unique key can be set on all fields.
HA_USE_AUTO_PARTITION
Set if the handler sets all tables to be partitioned by default.
*/
static uint partition_flags()
{
return HA_CAN_PARTITION;
}
static uint alter_table_flags(uint flags __attribute__((unused)))
{
return (HA_PARTITION_FUNCTION_SUPPORTED |
HA_FAST_CHANGE_PARTITION);
}
const uint32 ha_partition::NO_CURRENT_PART_ID= NOT_A_PARTITION_ID;
/*
Constructor method
SYNOPSIS
ha_partition()
table Table object
RETURN VALUE
NONE
*/
ha_partition::ha_partition(handlerton *hton, TABLE_SHARE *share)
:handler(hton, share)
{
DBUG_ENTER("ha_partition::ha_partition(table)");
init_alloc_root(&m_mem_root, 512, 512, MYF(0));
init_handler_variables();
DBUG_VOID_RETURN;
}
/*
Constructor method
SYNOPSIS
ha_partition()
part_info Partition info
RETURN VALUE
NONE
*/
ha_partition::ha_partition(handlerton *hton, partition_info *part_info)
:handler(hton, NULL)
{
DBUG_ENTER("ha_partition::ha_partition(part_info)");
DBUG_ASSERT(part_info);
init_alloc_root(&m_mem_root, 512, 512, MYF(0));
init_handler_variables();
m_part_info= part_info;
m_create_handler= TRUE;
m_is_sub_partitioned= m_part_info->is_sub_partitioned();
DBUG_VOID_RETURN;
}
/**
ha_partition constructor method used by ha_partition::clone()
@param hton Handlerton (partition_hton)
@param share Table share object
@param part_info_arg partition_info to use
@param clone_arg ha_partition to clone
@param clme_mem_root_arg MEM_ROOT to use
@return New partition handler
*/
ha_partition::ha_partition(handlerton *hton, TABLE_SHARE *share,
partition_info *part_info_arg,
ha_partition *clone_arg,
MEM_ROOT *clone_mem_root_arg)
:handler(hton, share)
{
DBUG_ENTER("ha_partition::ha_partition(clone)");
init_alloc_root(&m_mem_root, 512, 512, MYF(0));
init_handler_variables();
m_part_info= part_info_arg;
m_create_handler= TRUE;
m_is_sub_partitioned= m_part_info->is_sub_partitioned();
m_is_clone_of= clone_arg;
m_clone_mem_root= clone_mem_root_arg;
part_share= clone_arg->part_share;
m_tot_parts= clone_arg->m_tot_parts;
m_pkey_is_clustered= clone_arg->primary_key_is_clustered();
DBUG_VOID_RETURN;
}
/*
Initialize handler object
SYNOPSIS
init_handler_variables()
RETURN VALUE
NONE
*/
void ha_partition::init_handler_variables()
{
active_index= MAX_KEY;
m_mode= 0;
m_open_test_lock= 0;
m_file_buffer= NULL;
m_name_buffer_ptr= NULL;
m_engine_array= NULL;
m_connect_string= NULL;
m_file= NULL;
m_file_tot_parts= 0;
m_reorged_file= NULL;
m_new_file= NULL;
m_reorged_parts= 0;
m_added_file= NULL;
m_tot_parts= 0;
m_pkey_is_clustered= 0;
m_part_spec.start_part= NO_CURRENT_PART_ID;
m_scan_value= 2;
m_ref_length= 0;
m_part_spec.end_part= NO_CURRENT_PART_ID;
m_index_scan_type= partition_no_index_scan;
m_start_key.key= NULL;
m_start_key.length= 0;
m_myisam= FALSE;
m_innodb= FALSE;
m_extra_cache= FALSE;
m_extra_cache_size= 0;
m_extra_prepare_for_update= FALSE;
m_extra_cache_part_id= NO_CURRENT_PART_ID;
m_handler_status= handler_not_initialized;
m_part_field_array= NULL;
m_ordered_rec_buffer= NULL;
m_top_entry= NO_CURRENT_PART_ID;
m_rec_length= 0;
m_last_part= 0;
m_rec0= 0;
m_err_rec= NULL;
m_curr_key_info[0]= NULL;
m_curr_key_info[1]= NULL;
m_part_func_monotonicity_info= NON_MONOTONIC;
auto_increment_lock= FALSE;
auto_increment_safe_stmt_log_lock= FALSE;
/*
this allows blackhole to work properly
*/
m_num_locks= 0;
m_part_info= NULL;
m_create_handler= FALSE;
m_is_sub_partitioned= 0;
m_is_clone_of= NULL;
m_clone_mem_root= NULL;
part_share= NULL;
m_new_partitions_share_refs.empty();
m_part_ids_sorted_by_num_of_records= NULL;
#ifdef DONT_HAVE_TO_BE_INITALIZED
m_start_key.flag= 0;
m_ordered= TRUE;
#endif
}
const char *ha_partition::table_type() const
{
// we can do this since we only support a single engine type
return m_file[0]->table_type();
}
/*
Destructor method
SYNOPSIS
~ha_partition()
RETURN VALUE
NONE
*/
ha_partition::~ha_partition()
{
DBUG_ENTER("ha_partition::~ha_partition()");
if (m_new_partitions_share_refs.elements)
m_new_partitions_share_refs.delete_elements();
if (m_file != NULL)
{
uint i;
for (i= 0; i < m_tot_parts; i++)
delete m_file[i];
}
destroy_record_priority_queue();
my_free(m_part_ids_sorted_by_num_of_records);
if (m_added_file)
{
for (handler **ph= m_added_file; *ph; ph++)
delete (*ph);
}
clear_handler_file();
free_root(&m_mem_root, MYF(0));
DBUG_VOID_RETURN;
}
/*
Initialize partition handler object
SYNOPSIS
initialize_partition()
mem_root Allocate memory through this
RETURN VALUE
1 Error
0 Success
DESCRIPTION
The partition handler is only a layer on top of other engines. Thus it
can't really perform anything without the underlying handlers. Thus we
add this method as part of the allocation of a handler object.
1) Allocation of underlying handlers
If we have access to the partition info we will allocate one handler
instance for each partition.
2) Allocation without partition info
The cases where we don't have access to this information is when called
in preparation for delete_table and rename_table and in that case we
only need to set HA_FILE_BASED. In that case we will use the .par file
that contains information about the partitions and their engines and
the names of each partition.
3) Table flags initialisation
We need also to set table flags for the partition handler. This is not
static since it depends on what storage engines are used as underlying
handlers.
The table flags is set in this routine to simulate the behaviour of a
normal storage engine
The flag HA_FILE_BASED will be set independent of the underlying handlers
4) Index flags initialisation
When knowledge exists on the indexes it is also possible to initialize the
index flags. Again the index flags must be initialized by using the under-
lying handlers since this is storage engine dependent.
The flag HA_READ_ORDER will be reset for the time being to indicate no
ordered output is available from partition handler indexes. Later a merge
sort will be performed using the underlying handlers.
5) primary_key_is_clustered and has_transactions are
calculated here.
*/
bool ha_partition::initialize_partition(MEM_ROOT *mem_root)
{
handler **file_array, *file;
ulonglong check_table_flags;
DBUG_ENTER("ha_partition::initialize_partition");
if (m_create_handler)
{
m_tot_parts= m_part_info->get_tot_partitions();
DBUG_ASSERT(m_tot_parts > 0);
if (new_handlers_from_part_info(mem_root))
DBUG_RETURN(1);
}
else if (!table_share || !table_share->normalized_path.str)
{
/*
Called with dummy table share (delete, rename and alter table).
Don't need to set-up anything.
*/
DBUG_RETURN(0);
}
else if (get_from_handler_file(table_share->normalized_path.str,
mem_root, false))
{
my_error(ER_FAILED_READ_FROM_PAR_FILE, MYF(0));
DBUG_RETURN(1);
}
/*
We create all underlying table handlers here. We do it in this special
method to be able to report allocation errors.
Set up primary_key_is_clustered and
has_transactions since they are called often in all kinds of places,
other parameters are calculated on demand.
Verify that all partitions have the same table_flags.
*/
check_table_flags= m_file[0]->ha_table_flags();
m_pkey_is_clustered= TRUE;
file_array= m_file;
do
{
file= *file_array;
if (!file->primary_key_is_clustered())
m_pkey_is_clustered= FALSE;
if (check_table_flags != file->ha_table_flags())
{
my_error(ER_MIX_HANDLER_ERROR, MYF(0));
DBUG_RETURN(1);
}
} while (*(++file_array));
m_handler_status= handler_initialized;
DBUG_RETURN(0);
}
/****************************************************************************
MODULE meta data changes
****************************************************************************/
/*
Delete a table
SYNOPSIS
delete_table()
name Full path of table name
RETURN VALUE
>0 Error
0 Success
DESCRIPTION
Used to delete a table. By the time delete_table() has been called all
opened references to this table will have been closed (and your globally
shared references released. The variable name will just be the name of
the table. You will need to remove any files you have created at this
point.
If you do not implement this, the default delete_table() is called from
handler.cc and it will delete all files with the file extentions returned
by bas_ext().
Called from handler.cc by delete_table and ha_create_table(). Only used
during create if the table_flag HA_DROP_BEFORE_CREATE was specified for
the storage engine.
*/
int ha_partition::delete_table(const char *name)
{
DBUG_ENTER("ha_partition::delete_table");
DBUG_RETURN(del_ren_table(name, NULL));
}
/*
Rename a table
SYNOPSIS
rename_table()
from Full path of old table name
to Full path of new table name
RETURN VALUE
>0 Error
0 Success
DESCRIPTION
Renames a table from one name to another from alter table call.
If you do not implement this, the default rename_table() is called from
handler.cc and it will rename all files with the file extentions returned
by bas_ext().
Called from sql_table.cc by mysql_rename_table().
*/
int ha_partition::rename_table(const char *from, const char *to)
{
DBUG_ENTER("ha_partition::rename_table");
DBUG_RETURN(del_ren_table(from, to));
}
/*
Create the handler file (.par-file)
SYNOPSIS
create_partitioning_metadata()
name Full path of table name
create_info Create info generated for CREATE TABLE
RETURN VALUE
>0 Error
0 Success
DESCRIPTION
create_partitioning_metadata is called to create any handler specific files
before opening the file with openfrm to later call ::create on the
file object.
In the partition handler this is used to store the names of partitions
and types of engines in the partitions.
*/
int ha_partition::create_partitioning_metadata(const char *path,
const char *old_path,
int action_flag)
{
DBUG_ENTER("ha_partition::create_partitioning_metadata()");
/*
We need to update total number of parts since we might write the handler
file as part of a partition management command
*/
if (action_flag == CHF_DELETE_FLAG ||
action_flag == CHF_RENAME_FLAG)
{
char name[FN_REFLEN];
char old_name[FN_REFLEN];
strxmov(name, path, ha_par_ext, NullS);
strxmov(old_name, old_path, ha_par_ext, NullS);
if ((action_flag == CHF_DELETE_FLAG &&
mysql_file_delete(key_file_partition, name, MYF(MY_WME))) ||
(action_flag == CHF_RENAME_FLAG &&
mysql_file_rename(key_file_partition, old_name, name, MYF(MY_WME))))
{
DBUG_RETURN(TRUE);
}
}
else if (action_flag == CHF_CREATE_FLAG)
{
if (create_handler_file(path))
{
my_error(ER_CANT_CREATE_HANDLER_FILE, MYF(0));
DBUG_RETURN(1);
}
}
DBUG_RETURN(0);
}
/*
Create a partitioned table
SYNOPSIS
create()
name Full path of table name
table_arg Table object
create_info Create info generated for CREATE TABLE
RETURN VALUE
>0 Error
0 Success
DESCRIPTION
create() is called to create a table. The variable name will have the name
of the table. When create() is called you do not need to worry about
opening the table. Also, the FRM file will have already been created so
adjusting create_info will not do you any good. You can overwrite the frm
file at this point if you wish to change the table definition, but there
are no methods currently provided for doing that.
Called from handler.cc by ha_create_table().
*/
int ha_partition::create(const char *name, TABLE *table_arg,
HA_CREATE_INFO *create_info)
{
int error;
char name_buff[FN_REFLEN + 1], name_lc_buff[FN_REFLEN];
char *name_buffer_ptr;
const char *path;
uint i;
List_iterator_fast <partition_element> part_it(m_part_info->partitions);
partition_element *part_elem;
handler **file, **abort_file;
DBUG_ENTER("ha_partition::create");
DBUG_ASSERT(!fn_frm_ext(name));
/* Not allowed to create temporary partitioned tables */
if (create_info && create_info->tmp_table())
{
my_error(ER_PARTITION_NO_TEMPORARY, MYF(0));
DBUG_RETURN(TRUE);
}
if (get_from_handler_file(name, ha_thd()->mem_root, false))
DBUG_RETURN(TRUE);
DBUG_ASSERT(m_file_buffer);
DBUG_PRINT("enter", ("name: (%s)", name));
name_buffer_ptr= m_name_buffer_ptr;
file= m_file;
/*
Since ha_partition has HA_FILE_BASED, it must alter underlying table names
if they do not have HA_FILE_BASED and lower_case_table_names == 2.
See Bug#37402, for Mac OS X.
The appended #P#<partname>[#SP#<subpartname>] will remain in current case.
Using the first partitions handler, since mixing handlers is not allowed.
*/
path= get_canonical_filename(*file, name, name_lc_buff);
for (i= 0; i < m_part_info->num_parts; i++)
{
part_elem= part_it++;
if (m_is_sub_partitioned)
{
uint j;
List_iterator_fast <partition_element> sub_it(part_elem->subpartitions);
for (j= 0; j < m_part_info->num_subparts; j++)
{
part_elem= sub_it++;
if ((error= create_partition_name(name_buff, sizeof(name_buff), path,
name_buffer_ptr, NORMAL_PART_NAME, FALSE)))
goto create_error;
if ((error= set_up_table_before_create(table_arg, name_buff,
create_info, part_elem)) ||
((error= (*file)->ha_create(name_buff, table_arg, create_info))))
goto create_error;
name_buffer_ptr= strend(name_buffer_ptr) + 1;
file++;
}
}
else
{
if ((error= create_partition_name(name_buff, sizeof(name_buff), path,
name_buffer_ptr, NORMAL_PART_NAME, FALSE)))
goto create_error;
if ((error= set_up_table_before_create(table_arg, name_buff,
create_info, part_elem)) ||
((error= (*file)->ha_create(name_buff, table_arg, create_info))))
goto create_error;
name_buffer_ptr= strend(name_buffer_ptr) + 1;
file++;
}
}
DBUG_RETURN(0);
create_error:
name_buffer_ptr= m_name_buffer_ptr;
for (abort_file= file, file= m_file; file < abort_file; file++)
{
if (!create_partition_name(name_buff, sizeof(name_buff), path,
name_buffer_ptr, NORMAL_PART_NAME, FALSE))
(void) (*file)->ha_delete_table((const char*) name_buff);
name_buffer_ptr= strend(name_buffer_ptr) + 1;
}
handler::delete_table(name);
DBUG_RETURN(error);
}
/*
Drop partitions as part of ALTER TABLE of partitions
SYNOPSIS
drop_partitions()
path Complete path of db and table name
RETURN VALUE
>0 Failure
0 Success
DESCRIPTION
Use part_info object on handler object to deduce which partitions to
drop (each partition has a state attached to it)
*/
int ha_partition::drop_partitions(const char *path)
{
List_iterator<partition_element> part_it(m_part_info->partitions);
char part_name_buff[FN_REFLEN + 1];
uint num_parts= m_part_info->partitions.elements;
uint num_subparts= m_part_info->num_subparts;
uint i= 0;
uint name_variant;
int ret_error;
int error= 0;
DBUG_ENTER("ha_partition::drop_partitions");
/*
Assert that it works without HA_FILE_BASED and lower_case_table_name = 2.
We use m_file[0] as long as all partitions have the same storage engine.
*/
DBUG_ASSERT(!strcmp(path, get_canonical_filename(m_file[0], path,
part_name_buff)));
do
{
partition_element *part_elem= part_it++;
if (part_elem->part_state == PART_TO_BE_DROPPED)
{
handler *file;
/*
This part is to be dropped, meaning the part or all its subparts.
*/
name_variant= NORMAL_PART_NAME;
if (m_is_sub_partitioned)
{
List_iterator<partition_element> sub_it(part_elem->subpartitions);
uint j= 0, part;
do
{
partition_element *sub_elem= sub_it++;
part= i * num_subparts + j;
if ((ret_error= create_subpartition_name(part_name_buff,
sizeof(part_name_buff), path,
part_elem->partition_name,
sub_elem->partition_name, name_variant)))
error= ret_error;
file= m_file[part];
DBUG_PRINT("info", ("Drop subpartition %s", part_name_buff));
if ((ret_error= file->ha_delete_table(part_name_buff)))
error= ret_error;
if (deactivate_ddl_log_entry(sub_elem->log_entry->entry_pos))
error= 1;
} while (++j < num_subparts);
}
else
{
if ((ret_error= create_partition_name(part_name_buff,
sizeof(part_name_buff), path,
part_elem->partition_name, name_variant, TRUE)))
error= ret_error;
else
{
file= m_file[i];
DBUG_PRINT("info", ("Drop partition %s", part_name_buff));
if ((ret_error= file->ha_delete_table(part_name_buff)))
error= ret_error;
if (deactivate_ddl_log_entry(part_elem->log_entry->entry_pos))
error= 1;
}
}
if (part_elem->part_state == PART_IS_CHANGED)
part_elem->part_state= PART_NORMAL;
else
part_elem->part_state= PART_IS_DROPPED;
}
} while (++i < num_parts);
(void) sync_ddl_log();
DBUG_RETURN(error);
}
/*
Rename partitions as part of ALTER TABLE of partitions
SYNOPSIS
rename_partitions()
path Complete path of db and table name
RETURN VALUE
TRUE Failure
FALSE Success
DESCRIPTION
When reorganising partitions, adding hash partitions and coalescing
partitions it can be necessary to rename partitions while holding
an exclusive lock on the table.
Which partitions to rename is given by state of partitions found by the
partition info struct referenced from the handler object
*/
int ha_partition::rename_partitions(const char *path)
{
List_iterator<partition_element> part_it(m_part_info->partitions);
List_iterator<partition_element> temp_it(m_part_info->temp_partitions);
char part_name_buff[FN_REFLEN + 1];
char norm_name_buff[FN_REFLEN + 1];
uint num_parts= m_part_info->partitions.elements;
uint part_count= 0;
uint num_subparts= m_part_info->num_subparts;
uint i= 0;
uint j= 0;
int error= 0;
int ret_error;
uint temp_partitions= m_part_info->temp_partitions.elements;
handler *file;
partition_element *part_elem, *sub_elem;
DBUG_ENTER("ha_partition::rename_partitions");
/*
Assert that it works without HA_FILE_BASED and lower_case_table_name = 2.
We use m_file[0] as long as all partitions have the same storage engine.
*/
DBUG_ASSERT(!strcmp(path, get_canonical_filename(m_file[0], path,
norm_name_buff)));
DEBUG_SYNC(ha_thd(), "before_rename_partitions");
if (temp_partitions)
{
/*
These are the reorganised partitions that have already been copied.
We delete the partitions and log the delete by inactivating the
delete log entry in the table log. We only need to synchronise
these writes before moving to the next loop since there is no
interaction among reorganised partitions, they cannot have the
same name.
*/
do
{
part_elem= temp_it++;
if (m_is_sub_partitioned)
{
List_iterator<partition_element> sub_it(part_elem->subpartitions);
j= 0;
do
{
sub_elem= sub_it++;
file= m_reorged_file[part_count++];
if ((ret_error= create_subpartition_name(norm_name_buff,
sizeof(norm_name_buff), path,
part_elem->partition_name,
sub_elem->partition_name, NORMAL_PART_NAME)))
error= ret_error;
DBUG_PRINT("info", ("Delete subpartition %s", norm_name_buff));
if ((ret_error= file->ha_delete_table(norm_name_buff)))
error= ret_error;
else if (deactivate_ddl_log_entry(sub_elem->log_entry->entry_pos))
error= 1;
else
sub_elem->log_entry= NULL; /* Indicate success */
} while (++j < num_subparts);
}
else
{
file= m_reorged_file[part_count++];
if ((ret_error= create_partition_name(norm_name_buff,
sizeof(norm_name_buff), path,
part_elem->partition_name, NORMAL_PART_NAME, TRUE)))
error= ret_error;
else
{
DBUG_PRINT("info", ("Delete partition %s", norm_name_buff));
if ((ret_error= file->ha_delete_table(norm_name_buff)))
error= ret_error;
else if (deactivate_ddl_log_entry(part_elem->log_entry->entry_pos))
error= 1;
else
part_elem->log_entry= NULL; /* Indicate success */
}
}
} while (++i < temp_partitions);
(void) sync_ddl_log();
}
i= 0;
do
{
/*
When state is PART_IS_CHANGED it means that we have created a new
TEMP partition that is to be renamed to normal partition name and
we are to delete the old partition with currently the normal name.
We perform this operation by
1) Delete old partition with normal partition name
2) Signal this in table log entry
3) Synch table log to ensure we have consistency in crashes
4) Rename temporary partition name to normal partition name
5) Signal this to table log entry
It is not necessary to synch the last state since a new rename
should not corrupt things if there was no temporary partition.
The only other parts we need to cater for are new parts that
replace reorganised parts. The reorganised parts were deleted
by the code above that goes through the temp_partitions list.
Thus the synch above makes it safe to simply perform step 4 and 5
for those entries.
*/
part_elem= part_it++;
if (part_elem->part_state == PART_IS_CHANGED ||
part_elem->part_state == PART_TO_BE_DROPPED ||
(part_elem->part_state == PART_IS_ADDED && temp_partitions))
{
if (m_is_sub_partitioned)
{
List_iterator<partition_element> sub_it(part_elem->subpartitions);
uint part;
j= 0;
do
{
sub_elem= sub_it++;
part= i * num_subparts + j;
if ((ret_error= create_subpartition_name(norm_name_buff,
sizeof(norm_name_buff), path,
part_elem->partition_name,
sub_elem->partition_name, NORMAL_PART_NAME)))
error= ret_error;
if (part_elem->part_state == PART_IS_CHANGED)
{
file= m_reorged_file[part_count++];
DBUG_PRINT("info", ("Delete subpartition %s", norm_name_buff));
if ((ret_error= file->ha_delete_table(norm_name_buff)))
error= ret_error;
else if (deactivate_ddl_log_entry(sub_elem->log_entry->entry_pos))
error= 1;
(void) sync_ddl_log();
}
file= m_new_file[part];
if ((ret_error= create_subpartition_name(part_name_buff,
sizeof(part_name_buff), path,
part_elem->partition_name,
sub_elem->partition_name, TEMP_PART_NAME)))
error= ret_error;
DBUG_PRINT("info", ("Rename subpartition from %s to %s",
part_name_buff, norm_name_buff));
if ((ret_error= file->ha_rename_table(part_name_buff,
norm_name_buff)))
error= ret_error;
else if (deactivate_ddl_log_entry(sub_elem->log_entry->entry_pos))
error= 1;
else
sub_elem->log_entry= NULL;
} while (++j < num_subparts);
}
else
{
if ((ret_error= create_partition_name(norm_name_buff,
sizeof(norm_name_buff), path,
part_elem->partition_name, NORMAL_PART_NAME, TRUE)) ||
(ret_error= create_partition_name(part_name_buff,
sizeof(part_name_buff), path,
part_elem->partition_name, TEMP_PART_NAME, TRUE)))
error= ret_error;
else
{
if (part_elem->part_state == PART_IS_CHANGED)
{
file= m_reorged_file[part_count++];
DBUG_PRINT("info", ("Delete partition %s", norm_name_buff));
if ((ret_error= file->ha_delete_table(norm_name_buff)))
error= ret_error;
else if (deactivate_ddl_log_entry(part_elem->log_entry->entry_pos))
error= 1;
(void) sync_ddl_log();
}
file= m_new_file[i];
DBUG_PRINT("info", ("Rename partition from %s to %s",
part_name_buff, norm_name_buff));
if ((ret_error= file->ha_rename_table(part_name_buff,
norm_name_buff)))
error= ret_error;
else if (deactivate_ddl_log_entry(part_elem->log_entry->entry_pos))
error= 1;
else
part_elem->log_entry= NULL;
}
}
}
} while (++i < num_parts);
(void) sync_ddl_log();
DBUG_RETURN(error);
}
#define OPTIMIZE_PARTS 1
#define ANALYZE_PARTS 2
#define CHECK_PARTS 3
#define REPAIR_PARTS 4
#define ASSIGN_KEYCACHE_PARTS 5
#define PRELOAD_KEYS_PARTS 6
static const char *opt_op_name[]= {NULL,
"optimize", "analyze", "check", "repair",
"assign_to_keycache", "preload_keys"};
/*
Optimize table
SYNOPSIS
optimize()
thd Thread object
check_opt Check/analyze/repair/optimize options
RETURN VALUES
>0 Error
0 Success
*/
int ha_partition::optimize(THD *thd, HA_CHECK_OPT *check_opt)
{
DBUG_ENTER("ha_partition::optimize");
DBUG_RETURN(handle_opt_partitions(thd, check_opt, OPTIMIZE_PARTS));
}
/*
Analyze table
SYNOPSIS
analyze()
thd Thread object
check_opt Check/analyze/repair/optimize options
RETURN VALUES
>0 Error
0 Success
*/
int ha_partition::analyze(THD *thd, HA_CHECK_OPT *check_opt)
{
DBUG_ENTER("ha_partition::analyze");
DBUG_RETURN(handle_opt_partitions(thd, check_opt, ANALYZE_PARTS));
}
/*
Check table
SYNOPSIS
check()
thd Thread object
check_opt Check/analyze/repair/optimize options
RETURN VALUES
>0 Error
0 Success
*/
int ha_partition::check(THD *thd, HA_CHECK_OPT *check_opt)
{
DBUG_ENTER("ha_partition::check");
DBUG_RETURN(handle_opt_partitions(thd, check_opt, CHECK_PARTS));
}
/*
Repair table
SYNOPSIS
repair()
thd Thread object
check_opt Check/analyze/repair/optimize options
RETURN VALUES
>0 Error
0 Success
*/
int ha_partition::repair(THD *thd, HA_CHECK_OPT *check_opt)
{
DBUG_ENTER("ha_partition::repair");
int res= handle_opt_partitions(thd, check_opt, REPAIR_PARTS);
DBUG_RETURN(res);
}
/**
Assign to keycache
@param thd Thread object
@param check_opt Check/analyze/repair/optimize options
@return
@retval >0 Error
@retval 0 Success
*/
int ha_partition::assign_to_keycache(THD *thd, HA_CHECK_OPT *check_opt)
{
DBUG_ENTER("ha_partition::assign_to_keycache");
DBUG_RETURN(handle_opt_partitions(thd, check_opt, ASSIGN_KEYCACHE_PARTS));
}
/**
Preload to keycache
@param thd Thread object
@param check_opt Check/analyze/repair/optimize options
@return
@retval >0 Error
@retval 0 Success
*/
int ha_partition::preload_keys(THD *thd, HA_CHECK_OPT *check_opt)
{
DBUG_ENTER("ha_partition::preload_keys");
DBUG_RETURN(handle_opt_partitions(thd, check_opt, PRELOAD_KEYS_PARTS));
}
/*
Handle optimize/analyze/check/repair of one partition
SYNOPSIS
handle_opt_part()
thd Thread object
check_opt Options
file Handler object of partition
flag Optimize/Analyze/Check/Repair flag
RETURN VALUE
>0 Failure
0 Success
*/
int ha_partition::handle_opt_part(THD *thd, HA_CHECK_OPT *check_opt,
uint part_id, uint flag)
{
int error;
handler *file= m_file[part_id];
DBUG_ENTER("handle_opt_part");
DBUG_PRINT("enter", ("flag = %u", flag));
if (flag == OPTIMIZE_PARTS)
error= file->ha_optimize(thd, check_opt);
else if (flag == ANALYZE_PARTS)
error= file->ha_analyze(thd, check_opt);
else if (flag == CHECK_PARTS)
{
error= file->ha_check(thd, check_opt);
if (!error ||
error == HA_ADMIN_ALREADY_DONE ||
error == HA_ADMIN_NOT_IMPLEMENTED)
{
if (check_opt->flags & (T_MEDIUM | T_EXTEND))
error= check_misplaced_rows(part_id, false);
}
}
else if (flag == REPAIR_PARTS)
{
error= file->ha_repair(thd, check_opt);
if (!error ||
error == HA_ADMIN_ALREADY_DONE ||
error == HA_ADMIN_NOT_IMPLEMENTED)
{
if (check_opt->flags & (T_MEDIUM | T_EXTEND))
error= check_misplaced_rows(part_id, true);
}
}
else if (flag == ASSIGN_KEYCACHE_PARTS)
error= file->assign_to_keycache(thd, check_opt);
else if (flag == PRELOAD_KEYS_PARTS)
error= file->preload_keys(thd, check_opt);
else
{
DBUG_ASSERT(FALSE);
error= 1;
}
if (error == HA_ADMIN_ALREADY_DONE)
error= 0;
DBUG_RETURN(error);
}
/*
print a message row formatted for ANALYZE/CHECK/OPTIMIZE/REPAIR TABLE
(modelled after mi_check_print_msg)
TODO: move this into the handler, or rewrite mysql_admin_table.
*/
bool print_admin_msg(THD* thd, uint len,
const char* msg_type,
const char* db_name, String &table_name,
const char* op_name, const char *fmt, ...)
ATTRIBUTE_FORMAT(printf, 7, 8);
bool print_admin_msg(THD* thd, uint len,
const char* msg_type,
const char* db_name, String &table_name,
const char* op_name, const char *fmt, ...)
{
va_list args;
Protocol *protocol= thd->protocol;
uint length;
uint msg_length;
char name[NAME_LEN*2+2];
char *msgbuf;
bool error= true;
if (!(msgbuf= (char*) my_malloc(len, MYF(0))))
return true;
va_start(args, fmt);
msg_length= my_vsnprintf(msgbuf, len, fmt, args);
va_end(args);
if (msg_length >= (len - 1))
goto err;
msgbuf[len - 1] = 0; // healthy paranoia
if (!thd->vio_ok())
{
sql_print_error("%s", msgbuf);
goto err;
}
length=(uint) (strxmov(name, db_name, ".", table_name.c_ptr_safe(), NullS) - name);
/*
TODO: switch from protocol to push_warning here. The main reason we didn't
it yet is parallel repair, which threads have no THD object accessible via
current_thd.
Also we likely need to lock mutex here (in both cases with protocol and
push_warning).
*/
DBUG_PRINT("info",("print_admin_msg: %s, %s, %s, %s", name, op_name,
msg_type, msgbuf));
protocol->prepare_for_resend();
protocol->store(name, length, system_charset_info);
protocol->store(op_name, system_charset_info);
protocol->store(msg_type, system_charset_info);
protocol->store(msgbuf, msg_length, system_charset_info);
if (protocol->write())
{
sql_print_error("Failed on my_net_write, writing to stderr instead: %s\n",
msgbuf);
goto err;
}
error= false;
err:
my_free(msgbuf);
return error;
}
/*
Handle optimize/analyze/check/repair of partitions
SYNOPSIS
handle_opt_partitions()
thd Thread object
check_opt Options
flag Optimize/Analyze/Check/Repair flag
RETURN VALUE
>0 Failure
0 Success
*/
int ha_partition::handle_opt_partitions(THD *thd, HA_CHECK_OPT *check_opt,
uint flag)
{
List_iterator<partition_element> part_it(m_part_info->partitions);
uint num_parts= m_part_info->num_parts;
uint num_subparts= m_part_info->num_subparts;
uint i= 0;
int error;
DBUG_ENTER("ha_partition::handle_opt_partitions");
DBUG_PRINT("enter", ("flag= %u", flag));
do
{
partition_element *part_elem= part_it++;
/*
when ALTER TABLE <CMD> PARTITION ...
it should only do named partitions, otherwise all partitions
*/
if (!(thd->lex->alter_info.flags & Alter_info::ALTER_ADMIN_PARTITION) ||
part_elem->part_state == PART_ADMIN)
{
if (m_is_sub_partitioned)
{
List_iterator<partition_element> subpart_it(part_elem->subpartitions);
partition_element *sub_elem;
uint j= 0, part;
do
{
sub_elem= subpart_it++;
part= i * num_subparts + j;
DBUG_PRINT("info", ("Optimize subpartition %u (%s)",
part, sub_elem->partition_name));
if ((error= handle_opt_part(thd, check_opt, part, flag)))
{
/* print a line which partition the error belongs to */
if (error != HA_ADMIN_NOT_IMPLEMENTED &&
error != HA_ADMIN_ALREADY_DONE &&
error != HA_ADMIN_TRY_ALTER)
{
print_admin_msg(thd, MYSQL_ERRMSG_SIZE, "error",
table_share->db.str, table->alias,
opt_op_name[flag],
"Subpartition %s returned error",
sub_elem->partition_name);
}
/* reset part_state for the remaining partitions */
do
{
if (part_elem->part_state == PART_ADMIN)
part_elem->part_state= PART_NORMAL;
} while ((part_elem= part_it++));
DBUG_RETURN(error);
}
} while (++j < num_subparts);
}
else
{
DBUG_PRINT("info", ("Optimize partition %u (%s)", i,
part_elem->partition_name));
if ((error= handle_opt_part(thd, check_opt, i, flag)))
{
/* print a line which partition the error belongs to */
if (error != HA_ADMIN_NOT_IMPLEMENTED &&
error != HA_ADMIN_ALREADY_DONE &&
error != HA_ADMIN_TRY_ALTER)
{
print_admin_msg(thd, MYSQL_ERRMSG_SIZE, "error",
table_share->db.str, table->alias,
opt_op_name[flag], "Partition %s returned error",
part_elem->partition_name);
}
/* reset part_state for the remaining partitions */
do
{
if (part_elem->part_state == PART_ADMIN)
part_elem->part_state= PART_NORMAL;
} while ((part_elem= part_it++));
DBUG_RETURN(error);
}
}
part_elem->part_state= PART_NORMAL;
}
} while (++i < num_parts);
DBUG_RETURN(FALSE);
}
/**
@brief Check and repair the table if neccesary
@param thd Thread object
@retval TRUE Error/Not supported
@retval FALSE Success
@note Called if open_table_from_share fails and ::is_crashed().
*/
bool ha_partition::check_and_repair(THD *thd)
{
handler **file= m_file;
DBUG_ENTER("ha_partition::check_and_repair");
do
{
if ((*file)->ha_check_and_repair(thd))
DBUG_RETURN(TRUE);
} while (*(++file));
DBUG_RETURN(FALSE);
}
/**
@breif Check if the table can be automatically repaired
@retval TRUE Can be auto repaired
@retval FALSE Cannot be auto repaired
*/
bool ha_partition::auto_repair(int error) const
{
DBUG_ENTER("ha_partition::auto_repair");
/*
As long as we only support one storage engine per table,
we can use the first partition for this function.
*/
DBUG_RETURN(m_file[0]->auto_repair(error));
}
/**
@breif Check if the table is crashed
@retval TRUE Crashed
@retval FALSE Not crashed
*/
bool ha_partition::is_crashed() const
{
handler **file= m_file;
DBUG_ENTER("ha_partition::is_crashed");
do
{
if ((*file)->is_crashed())
DBUG_RETURN(TRUE);
} while (*(++file));
DBUG_RETURN(FALSE);
}
/*
Prepare by creating a new partition
SYNOPSIS
prepare_new_partition()
table Table object
create_info Create info from CREATE TABLE
file Handler object of new partition
part_name partition name
RETURN VALUE
>0 Error
0 Success
*/
int ha_partition::prepare_new_partition(TABLE *tbl,
HA_CREATE_INFO *create_info,
handler *file, const char *part_name,
partition_element *p_elem,
uint disable_non_uniq_indexes)
{
int error;
DBUG_ENTER("prepare_new_partition");
/*
This call to set_up_table_before_create() is done for an alter table.
So this may be the second time around for this partition_element,
depending on how many partitions and subpartitions there were before,
and how many there are now.
The first time, on the CREATE, data_file_name and index_file_name
came from the parser. They did not have the file name attached to
the end. But if this partition is less than the total number of
previous partitions, it's data_file_name has the filename attached.
So we need to take the partition filename off if it exists.
That file name may be different from part_name, which will be
attached in append_file_to_dir().
*/
truncate_partition_filename((char*) p_elem->data_file_name);
truncate_partition_filename((char*) p_elem->index_file_name);
if ((error= set_up_table_before_create(tbl, part_name, create_info, p_elem)))
goto error_create;
tbl->s->connect_string = p_elem->connect_string;
if ((error= file->ha_create(part_name, tbl, create_info)))
{
/*
Added for safety, InnoDB reports HA_ERR_FOUND_DUPP_KEY
if the table/partition already exists.
If we return that error code, then print_error would try to
get_dup_key on a non-existing partition.
So return a more reasonable error code.
*/
if (error == HA_ERR_FOUND_DUPP_KEY)
error= HA_ERR_TABLE_EXIST;
goto error_create;
}
DBUG_PRINT("info", ("partition %s created", part_name));
if ((error= file->ha_open(tbl, part_name, m_mode,
m_open_test_lock | HA_OPEN_NO_PSI_CALL)))
goto error_open;
DBUG_PRINT("info", ("partition %s opened", part_name));
/*
Note: if you plan to add another call that may return failure,
better to do it before external_lock() as cleanup_new_partition()
assumes that external_lock() is last call that may fail here.
Otherwise see description for cleanup_new_partition().
*/
if ((error= file->ha_external_lock(ha_thd(), F_WRLCK)))
goto error_external_lock;
DBUG_PRINT("info", ("partition %s external locked", part_name));
if (disable_non_uniq_indexes)
file->ha_disable_indexes(HA_KEY_SWITCH_NONUNIQ_SAVE);
DBUG_RETURN(0);
error_external_lock:
(void) file->ha_close();
error_open:
(void) file->ha_delete_table(part_name);
error_create:
DBUG_RETURN(error);
}
/*
Cleanup by removing all created partitions after error
SYNOPSIS
cleanup_new_partition()
part_count Number of partitions to remove
RETURN VALUE
NONE
DESCRIPTION
This function is called immediately after prepare_new_partition() in
case the latter fails.
In prepare_new_partition() last call that may return failure is
external_lock(). That means if prepare_new_partition() fails,
partition does not have external lock. Thus no need to call
external_lock(F_UNLCK) here.
TODO:
We must ensure that in the case that we get an error during the process
that we call external_lock with F_UNLCK, close the table and delete the
table in the case where we have been successful with prepare_handler.
We solve this by keeping an array of successful calls to prepare_handler
which can then be used to undo the call.
*/
void ha_partition::cleanup_new_partition(uint part_count)
{
DBUG_ENTER("ha_partition::cleanup_new_partition");
if (m_added_file)
{
THD *thd= ha_thd();
handler **file= m_added_file;
while ((part_count > 0) && (*file))
{
(*file)->ha_external_lock(thd, F_UNLCK);
(*file)->ha_close();
/* Leave the (*file)->ha_delete_table(part_name) to the ddl-log */
file++;
part_count--;
}
m_added_file= NULL;
}
DBUG_VOID_RETURN;
}
/*
Implement the partition changes defined by ALTER TABLE of partitions
SYNOPSIS
change_partitions()
create_info HA_CREATE_INFO object describing all
fields and indexes in table
path Complete path of db and table name
out: copied Output parameter where number of copied
records are added
out: deleted Output parameter where number of deleted
records are added
pack_frm_data Reference to packed frm file
pack_frm_len Length of packed frm file
RETURN VALUE
>0 Failure
0 Success
DESCRIPTION
Add and copy if needed a number of partitions, during this operation
no other operation is ongoing in the server. This is used by
ADD PARTITION all types as well as by REORGANIZE PARTITION. For
one-phased implementations it is used also by DROP and COALESCE
PARTITIONs.
One-phased implementation needs the new frm file, other handlers will
get zero length and a NULL reference here.
*/
int ha_partition::change_partitions(HA_CREATE_INFO *create_info,
const char *path,
ulonglong * const copied,
ulonglong * const deleted,
const uchar *pack_frm_data
__attribute__((unused)),
size_t pack_frm_len
__attribute__((unused)))
{
List_iterator<partition_element> part_it(m_part_info->partitions);
List_iterator <partition_element> t_it(m_part_info->temp_partitions);
char part_name_buff[FN_REFLEN + 1];
uint num_parts= m_part_info->partitions.elements;
uint num_subparts= m_part_info->num_subparts;
uint i= 0;
uint num_remain_partitions, part_count, orig_count;
handler **new_file_array;
int error= 1;
bool first;
uint temp_partitions= m_part_info->temp_partitions.elements;
THD *thd= ha_thd();
DBUG_ENTER("ha_partition::change_partitions");
/*
Assert that it works without HA_FILE_BASED and lower_case_table_name = 2.
We use m_file[0] as long as all partitions have the same storage engine.
*/
DBUG_ASSERT(!strcmp(path, get_canonical_filename(m_file[0], path,
part_name_buff)));
m_reorged_parts= 0;
if (!m_part_info->is_sub_partitioned())
num_subparts= 1;
/*
Step 1:
Calculate number of reorganised partitions and allocate space for
their handler references.
*/
if (temp_partitions)
{
m_reorged_parts= temp_partitions * num_subparts;
}
else
{
do
{
partition_element *part_elem= part_it++;
if (part_elem->part_state == PART_CHANGED ||
part_elem->part_state == PART_REORGED_DROPPED)
{
m_reorged_parts+= num_subparts;
}
} while (++i < num_parts);
}
if (m_reorged_parts &&
!(m_reorged_file= (handler**) thd->calloc(sizeof(handler*)*
(m_reorged_parts + 1))))
{
mem_alloc_error(sizeof(handler*)*(m_reorged_parts+1));
DBUG_RETURN(HA_ERR_OUT_OF_MEM);
}
/*
Step 2:
Calculate number of partitions after change and allocate space for
their handler references.
*/
num_remain_partitions= 0;
if (temp_partitions)
{
num_remain_partitions= num_parts * num_subparts;
}
else
{
part_it.rewind();
i= 0;
do
{
partition_element *part_elem= part_it++;
if (part_elem->part_state == PART_NORMAL ||
part_elem->part_state == PART_TO_BE_ADDED ||
part_elem->part_state == PART_CHANGED)
{
num_remain_partitions+= num_subparts;
}
} while (++i < num_parts);
}
if (!(new_file_array= ((handler**)
thd->calloc(sizeof(handler*)*
(2*(num_remain_partitions + 1))))))
{
mem_alloc_error(sizeof(handler*)*2*(num_remain_partitions+1));
DBUG_RETURN(HA_ERR_OUT_OF_MEM);
}
m_added_file= &new_file_array[num_remain_partitions + 1];
/*
Step 3:
Fill m_reorged_file with handler references and NULL at the end
*/
if (m_reorged_parts)
{
i= 0;
part_count= 0;
first= TRUE;
part_it.rewind();
do
{
partition_element *part_elem= part_it++;
if (part_elem->part_state == PART_CHANGED ||
part_elem->part_state == PART_REORGED_DROPPED)
{
memcpy((void*)&m_reorged_file[part_count],
(void*)&m_file[i*num_subparts],
sizeof(handler*)*num_subparts);
part_count+= num_subparts;
}
else if (first && temp_partitions &&
part_elem->part_state == PART_TO_BE_ADDED)
{
/*
When doing an ALTER TABLE REORGANIZE PARTITION a number of
partitions is to be reorganised into a set of new partitions.
The reorganised partitions are in this case in the temp_partitions
list. We copy all of them in one batch and thus we only do this
until we find the first partition with state PART_TO_BE_ADDED
since this is where the new partitions go in and where the old
ones used to be.
*/
first= FALSE;
DBUG_ASSERT(((i*num_subparts) + m_reorged_parts) <= m_file_tot_parts);
memcpy((void*)m_reorged_file, &m_file[i*num_subparts],
sizeof(handler*)*m_reorged_parts);
}
} while (++i < num_parts);
}
/*
Step 4:
Fill new_array_file with handler references. Create the handlers if
needed.
*/
i= 0;
part_count= 0;
orig_count= 0;
first= TRUE;
part_it.rewind();
do
{
partition_element *part_elem= part_it++;
if (part_elem->part_state == PART_NORMAL)
{
DBUG_ASSERT(orig_count + num_subparts <= m_file_tot_parts);
memcpy((void*)&new_file_array[part_count], (void*)&m_file[orig_count],
sizeof(handler*)*num_subparts);
part_count+= num_subparts;
orig_count+= num_subparts;
}
else if (part_elem->part_state == PART_CHANGED ||
part_elem->part_state == PART_TO_BE_ADDED)
{
uint j= 0;
Parts_share_refs *p_share_refs;
/*
The Handler_shares for each partition's handler can be allocated
within this handler, since there will not be any more instances of the
new partitions, until the table is reopened after the ALTER succeeded.
*/
p_share_refs= new Parts_share_refs;
if (!p_share_refs)
DBUG_RETURN(HA_ERR_OUT_OF_MEM);
if (p_share_refs->init(num_subparts))
DBUG_RETURN(HA_ERR_OUT_OF_MEM);
if (m_new_partitions_share_refs.push_back(p_share_refs, thd->mem_root))
DBUG_RETURN(HA_ERR_OUT_OF_MEM);
do
{
handler **new_file= &new_file_array[part_count++];
if (!(*new_file=
get_new_handler(table->s,
thd->mem_root,
part_elem->engine_type)))
{
mem_alloc_error(sizeof(handler));
DBUG_RETURN(HA_ERR_OUT_OF_MEM);
}
if ((*new_file)->set_ha_share_ref(&p_share_refs->ha_shares[j]))
{
DBUG_RETURN(HA_ERR_OUT_OF_MEM);
}
} while (++j < num_subparts);
if (part_elem->part_state == PART_CHANGED)
orig_count+= num_subparts;
else if (temp_partitions && first)
{
orig_count+= (num_subparts * temp_partitions);
first= FALSE;
}
}
} while (++i < num_parts);
first= FALSE;
/*
Step 5:
Create the new partitions and also open, lock and call external_lock
on them to prepare them for copy phase and also for later close
calls
*/
/*
Before creating new partitions check whether indexes are disabled
in the partitions.
*/
uint disable_non_uniq_indexes = indexes_are_disabled();
i= 0;
part_count= 0;
part_it.rewind();
do
{
partition_element *part_elem= part_it++;
if (part_elem->part_state == PART_TO_BE_ADDED ||
part_elem->part_state == PART_CHANGED)
{
/*
A new partition needs to be created PART_TO_BE_ADDED means an
entirely new partition and PART_CHANGED means a changed partition
that will still exist with either more or less data in it.
*/
uint name_variant= NORMAL_PART_NAME;
if (part_elem->part_state == PART_CHANGED ||
(part_elem->part_state == PART_TO_BE_ADDED && temp_partitions))
name_variant= TEMP_PART_NAME;
if (m_part_info->is_sub_partitioned())
{
List_iterator<partition_element> sub_it(part_elem->subpartitions);
uint j= 0, part;
do
{
partition_element *sub_elem= sub_it++;
if ((error= create_subpartition_name(part_name_buff,
sizeof(part_name_buff), path,
part_elem->partition_name, sub_elem->partition_name,
name_variant)))
{
cleanup_new_partition(part_count);
DBUG_RETURN(error);
}
part= i * num_subparts + j;
DBUG_PRINT("info", ("Add subpartition %s", part_name_buff));
if ((error= prepare_new_partition(table, create_info,
new_file_array[part],
(const char *)part_name_buff,
sub_elem,
disable_non_uniq_indexes)))
{
cleanup_new_partition(part_count);
DBUG_RETURN(error);
}
m_added_file[part_count++]= new_file_array[part];
} while (++j < num_subparts);
}
else
{
if ((error= create_partition_name(part_name_buff,
sizeof(part_name_buff), path, part_elem->partition_name,
name_variant, TRUE)))
{
cleanup_new_partition(part_count);
DBUG_RETURN(error);
}
DBUG_PRINT("info", ("Add partition %s", part_name_buff));
if ((error= prepare_new_partition(table, create_info,
new_file_array[i],
(const char *)part_name_buff,
part_elem,
disable_non_uniq_indexes)))
{
cleanup_new_partition(part_count);
DBUG_RETURN(error);
}
m_added_file[part_count++]= new_file_array[i];
}
}
} while (++i < num_parts);
/*
Step 6:
State update to prepare for next write of the frm file.
*/
i= 0;
part_it.rewind();
do
{
partition_element *part_elem= part_it++;
if (part_elem->part_state == PART_TO_BE_ADDED)
part_elem->part_state= PART_IS_ADDED;
else if (part_elem->part_state == PART_CHANGED)
part_elem->part_state= PART_IS_CHANGED;
else if (part_elem->part_state == PART_REORGED_DROPPED)
part_elem->part_state= PART_TO_BE_DROPPED;
} while (++i < num_parts);
for (i= 0; i < temp_partitions; i++)
{
partition_element *part_elem= t_it++;
DBUG_ASSERT(part_elem->part_state == PART_TO_BE_REORGED);
part_elem->part_state= PART_TO_BE_DROPPED;
}
m_new_file= new_file_array;
if ((error= copy_partitions(copied, deleted)))
{
/*
Close and unlock the new temporary partitions.
They will later be deleted through the ddl-log.
*/
cleanup_new_partition(part_count);
}
DBUG_RETURN(error);
}
/*
Copy partitions as part of ALTER TABLE of partitions
SYNOPSIS
copy_partitions()
out:copied Number of records copied
out:deleted Number of records deleted
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
change_partitions has done all the preparations, now it is time to
actually copy the data from the reorganised partitions to the new
partitions.
*/
int ha_partition::copy_partitions(ulonglong * const copied,
ulonglong * const deleted)
{
uint reorg_part= 0;
int result= 0;
longlong func_value;
DBUG_ENTER("ha_partition::copy_partitions");
if (m_part_info->linear_hash_ind)
{
if (m_part_info->part_type == HASH_PARTITION)
set_linear_hash_mask(m_part_info, m_part_info->num_parts);
else
set_linear_hash_mask(m_part_info, m_part_info->num_subparts);
}
while (reorg_part < m_reorged_parts)
{
handler *file= m_reorged_file[reorg_part];
uint32 new_part;
late_extra_cache(reorg_part);
if ((result= file->ha_rnd_init_with_error(1)))
goto init_error;
while (TRUE)
{
if ((result= file->ha_rnd_next(m_rec0)))
{
if (result == HA_ERR_RECORD_DELETED)
continue; //Probably MyISAM
if (result != HA_ERR_END_OF_FILE)
goto error;
/*
End-of-file reached, break out to continue with next partition or
end the copy process.
*/
break;
}
/* Found record to insert into new handler */
if (m_part_info->get_partition_id(m_part_info, &new_part,
&func_value))
{
/*
This record is in the original table but will not be in the new
table since it doesn't fit into any partition any longer due to
changed partitioning ranges or list values.
*/
(*deleted)++;
}
else
{
THD *thd= ha_thd();
/* Copy record to new handler */
(*copied)++;
tmp_disable_binlog(thd); /* Do not replicate the low-level changes. */
result= m_new_file[new_part]->ha_write_row(m_rec0);
reenable_binlog(thd);
if (result)
goto error;
}
}
late_extra_no_cache(reorg_part);
file->ha_rnd_end();
reorg_part++;
}
DBUG_RETURN(FALSE);
error:
m_reorged_file[reorg_part]->ha_rnd_end();
init_error:
DBUG_RETURN(result);
}
/*
Update create info as part of ALTER TABLE
SYNOPSIS
update_create_info()
create_info Create info from ALTER TABLE
RETURN VALUE
NONE
DESCRIPTION
Forward this handler call to the storage engine foreach
partition handler. The data_file_name for each partition may
need to be reset if the tablespace was moved. Use a dummy
HA_CREATE_INFO structure and transfer necessary data.
*/
void ha_partition::update_create_info(HA_CREATE_INFO *create_info)
{
DBUG_ENTER("ha_partition::update_create_info");
/*
Fix for bug#38751, some engines needs info-calls in ALTER.
Archive need this since it flushes in ::info.
HA_STATUS_AUTO is optimized so it will not always be forwarded
to all partitions, but HA_STATUS_VARIABLE will.
*/
info(HA_STATUS_VARIABLE);
info(HA_STATUS_AUTO);
if (!(create_info->used_fields & HA_CREATE_USED_AUTO))
create_info->auto_increment_value= stats.auto_increment_value;
/*
DATA DIRECTORY and INDEX DIRECTORY are never applied to the whole
partitioned table, only its parts.
*/
my_bool from_alter = (create_info->data_file_name == (const char*) -1);
create_info->data_file_name= create_info->index_file_name = NULL;
create_info->connect_string= null_clex_str;
/*
We do not need to update the individual partition DATA DIRECTORY settings
since they can be changed by ALTER TABLE ... REORGANIZE PARTITIONS.
*/
if (from_alter)
DBUG_VOID_RETURN;
/*
send Handler::update_create_info() to the storage engine for each
partition that currently has a handler object. Using a dummy
HA_CREATE_INFO structure to collect DATA and INDEX DIRECTORYs.
*/
List_iterator<partition_element> part_it(m_part_info->partitions);
partition_element *part_elem, *sub_elem;
uint num_subparts= m_part_info->num_subparts;
uint num_parts = num_subparts ? m_file_tot_parts / num_subparts
: m_file_tot_parts;
HA_CREATE_INFO dummy_info;
memset(&dummy_info, 0, sizeof(dummy_info));
/*
Since update_create_info() can be called from mysql_prepare_alter_table()
when not all handlers are set up, we look for that condition first.
If all handlers are not available, do not call update_create_info for any.
*/
uint i, j, part;
for (i= 0; i < num_parts; i++)
{
part_elem= part_it++;
if (!part_elem)
DBUG_VOID_RETURN;
if (m_is_sub_partitioned)
{
List_iterator<partition_element> subpart_it(part_elem->subpartitions);
for (j= 0; j < num_subparts; j++)
{
sub_elem= subpart_it++;
if (!sub_elem)
DBUG_VOID_RETURN;
part= i * num_subparts + j;
if (part >= m_file_tot_parts || !m_file[part])
DBUG_VOID_RETURN;
}
}
else
{
if (!m_file[i])
DBUG_VOID_RETURN;
}
}
part_it.rewind();
for (i= 0; i < num_parts; i++)
{
part_elem= part_it++;
DBUG_ASSERT(part_elem);
if (m_is_sub_partitioned)
{
List_iterator<partition_element> subpart_it(part_elem->subpartitions);
for (j= 0; j < num_subparts; j++)
{
sub_elem= subpart_it++;
DBUG_ASSERT(sub_elem);
part= i * num_subparts + j;
DBUG_ASSERT(part < m_file_tot_parts && m_file[part]);
if (ha_legacy_type(m_file[part]->ht) == DB_TYPE_INNODB)
{
dummy_info.data_file_name= dummy_info.index_file_name = NULL;
m_file[part]->update_create_info(&dummy_info);
if (dummy_info.data_file_name || sub_elem->data_file_name)
{
sub_elem->data_file_name = (char*) dummy_info.data_file_name;
}
if (dummy_info.index_file_name || sub_elem->index_file_name)
{
sub_elem->index_file_name = (char*) dummy_info.index_file_name;
}
}
}
}
else
{
DBUG_ASSERT(m_file[i]);
if (ha_legacy_type(m_file[i]->ht) == DB_TYPE_INNODB)
{
dummy_info.data_file_name= dummy_info.index_file_name= NULL;
m_file[i]->update_create_info(&dummy_info);
if (dummy_info.data_file_name || part_elem->data_file_name)
{
part_elem->data_file_name = (char*) dummy_info.data_file_name;
}
if (dummy_info.index_file_name || part_elem->index_file_name)
{
part_elem->index_file_name = (char*) dummy_info.index_file_name;
}
}
}
}
DBUG_VOID_RETURN;
}
/**
Change the internal TABLE_SHARE pointer
@param table_arg TABLE object
@param share New share to use
@note Is used in error handling in ha_delete_table.
All handlers should exist (lock_partitions should not be used)
*/
void ha_partition::change_table_ptr(TABLE *table_arg, TABLE_SHARE *share)
{
handler **file_array;
table= table_arg;
table_share= share;
/*
m_file can be NULL when using an old cached table in DROP TABLE, when the
table just has REMOVED PARTITIONING, see Bug#42438
*/
if (m_file)
{
file_array= m_file;
DBUG_ASSERT(*file_array);
do
{
(*file_array)->change_table_ptr(table_arg, share);
} while (*(++file_array));
}
if (m_added_file && m_added_file[0])
{
/* if in middle of a drop/rename etc */
file_array= m_added_file;
do
{
(*file_array)->change_table_ptr(table_arg, share);
} while (*(++file_array));
}
}
/*
Change comments specific to handler
SYNOPSIS
update_table_comment()
comment Original comment
RETURN VALUE
new comment
DESCRIPTION
No comment changes so far
*/
char *ha_partition::update_table_comment(const char *comment)
{
return (char*) comment; /* Nothing to change */
}
/**
Handle delete and rename table
@param from Full path of old table
@param to Full path of new table
@return Operation status
@retval >0 Error
@retval 0 Success
@note Common routine to handle delete_table and rename_table.
The routine uses the partition handler file to get the
names of the partition instances. Both these routines
are called after creating the handler without table
object and thus the file is needed to discover the
names of the partitions and the underlying storage engines.
*/
uint ha_partition::del_ren_table(const char *from, const char *to)
{
int save_error= 0;
int error;
char from_buff[FN_REFLEN + 1], to_buff[FN_REFLEN + 1],
from_lc_buff[FN_REFLEN], to_lc_buff[FN_REFLEN];
char *name_buffer_ptr;
const char *from_path;
const char *to_path= NULL;
uint i;
handler **file, **abort_file;
DBUG_ENTER("ha_partition::del_ren_table");
if (get_from_handler_file(from, ha_thd()->mem_root, false))
DBUG_RETURN(TRUE);
DBUG_ASSERT(m_file_buffer);
DBUG_PRINT("enter", ("from: (%s) to: (%s)", from, to ? to : "(nil)"));
name_buffer_ptr= m_name_buffer_ptr;
file= m_file;
if (to == NULL)
{
/*
Delete table, start by delete the .par file. If error, break, otherwise
delete as much as possible.
*/
if ((error= handler::delete_table(from)))
DBUG_RETURN(error);
}
/*
Since ha_partition has HA_FILE_BASED, it must alter underlying table names
if they do not have HA_FILE_BASED and lower_case_table_names == 2.
See Bug#37402, for Mac OS X.
The appended #P#<partname>[#SP#<subpartname>] will remain in current case.
Using the first partitions handler, since mixing handlers is not allowed.
*/
from_path= get_canonical_filename(*file, from, from_lc_buff);
if (to != NULL)
to_path= get_canonical_filename(*file, to, to_lc_buff);
i= 0;
do
{
if ((error= create_partition_name(from_buff, sizeof(from_buff), from_path,
name_buffer_ptr, NORMAL_PART_NAME, FALSE)))
goto rename_error;
if (to != NULL)
{ // Rename branch
if ((error= create_partition_name(to_buff, sizeof(to_buff), to_path,
name_buffer_ptr, NORMAL_PART_NAME, FALSE)))
goto rename_error;
error= (*file)->ha_rename_table(from_buff, to_buff);
if (error)
goto rename_error;
}
else // delete branch
{
error= (*file)->ha_delete_table(from_buff);
}
name_buffer_ptr= strend(name_buffer_ptr) + 1;
if (error)
save_error= error;
i++;
} while (*(++file));
if (to != NULL)
{
if ((error= handler::rename_table(from, to)))
{
/* Try to revert everything, ignore errors */
(void) handler::rename_table(to, from);
goto rename_error;
}
}
DBUG_RETURN(save_error);
rename_error:
name_buffer_ptr= m_name_buffer_ptr;
for (abort_file= file, file= m_file; file < abort_file; file++)
{
/* Revert the rename, back from 'to' to the original 'from' */
if (!create_partition_name(from_buff, sizeof(from_buff), from_path,
name_buffer_ptr, NORMAL_PART_NAME, FALSE) &&
!create_partition_name(to_buff, sizeof(to_buff), to_path,
name_buffer_ptr, NORMAL_PART_NAME, FALSE))
{
/* Ignore error here */
(void) (*file)->ha_rename_table(to_buff, from_buff);
}
name_buffer_ptr= strend(name_buffer_ptr) + 1;
}
DBUG_RETURN(error);
}
uint ha_partition::count_query_cache_dependant_tables(uint8 *tables_type)
{
DBUG_ENTER("ha_partition::count_query_cache_dependant_tables");
/* Here we rely on the fact that all tables are of the same type */
uint8 type= m_file[0]->table_cache_type();
(*tables_type)|= type;
DBUG_PRINT("info", ("cnt: %u", (uint)m_tot_parts));
/*
We need save underlying tables only for HA_CACHE_TBL_ASKTRANSACT:
HA_CACHE_TBL_NONTRANSACT - because all changes goes through partition table
HA_CACHE_TBL_NOCACHE - because will not be cached
HA_CACHE_TBL_TRANSACT - QC need to know that such type present
*/
DBUG_RETURN(type == HA_CACHE_TBL_ASKTRANSACT ? m_tot_parts : 0);
}
my_bool ha_partition::
reg_query_cache_dependant_table(THD *thd,
char *engine_key, uint engine_key_len,
char *cache_key, uint cache_key_len,
uint8 type,
Query_cache *cache,
Query_cache_block_table **block_table,
handler *file,
uint *n)
{
DBUG_ENTER("ha_partition::reg_query_cache_dependant_table");
qc_engine_callback engine_callback;
ulonglong engine_data;
/* ask undelying engine */
if (!file->register_query_cache_table(thd, engine_key,
engine_key_len,
&engine_callback,
&engine_data))
{
DBUG_PRINT("qcache", ("Handler does not allow caching for %.*s",
engine_key_len, engine_key));
/*
As this can change from call to call, don't reset set
thd->lex->safe_to_cache_query
*/
thd->query_cache_is_applicable= 0; // Query can't be cached
DBUG_RETURN(TRUE);
}
(++(*block_table))->n= ++(*n);
if (!cache->insert_table(thd, cache_key_len,
cache_key, (*block_table),
table_share->db.length,
(uint8) (cache_key_len -
table_share->table_cache_key.length),
type,
engine_callback, engine_data,
FALSE))
DBUG_RETURN(TRUE);
DBUG_RETURN(FALSE);
}
my_bool ha_partition::
register_query_cache_dependant_tables(THD *thd,
Query_cache *cache,
Query_cache_block_table **block_table,
uint *n)
{
char *engine_key_end, *query_cache_key_end;
uint i;
uint num_parts= m_part_info->num_parts;
uint num_subparts= m_part_info->num_subparts;
int diff_length;
List_iterator<partition_element> part_it(m_part_info->partitions);
char engine_key[FN_REFLEN], query_cache_key[FN_REFLEN];
DBUG_ENTER("ha_partition::register_query_cache_dependant_tables");
/* see ha_partition::count_query_cache_dependant_tables */
if (m_file[0]->table_cache_type() != HA_CACHE_TBL_ASKTRANSACT)
DBUG_RETURN(FALSE); // nothing to register
/* prepare static part of the key */
memcpy(engine_key, table_share->normalized_path.str,
table_share->normalized_path.length);
memcpy(query_cache_key, table_share->table_cache_key.str,
table_share->table_cache_key.length);
diff_length= ((int) table_share->table_cache_key.length -
(int) table_share->normalized_path.length -1);
engine_key_end= engine_key + table_share->normalized_path.length;
query_cache_key_end= query_cache_key + table_share->table_cache_key.length -1;
engine_key_end[0]= engine_key_end[2]= query_cache_key_end[0]=
query_cache_key_end[2]= '#';
query_cache_key_end[1]= engine_key_end[1]= 'P';
engine_key_end+= 3;
query_cache_key_end+= 3;
i= 0;
do
{
partition_element *part_elem= part_it++;
char *engine_pos= strmov(engine_key_end, part_elem->partition_name);
if (m_is_sub_partitioned)
{
List_iterator<partition_element> subpart_it(part_elem->subpartitions);
partition_element *sub_elem;
uint j= 0, part;
engine_pos[0]= engine_pos[3]= '#';
engine_pos[1]= 'S';
engine_pos[2]= 'P';
engine_pos += 4;
do
{
char *end;
uint length;
sub_elem= subpart_it++;
part= i * num_subparts + j;
/* we store the end \0 as part of the key */
end= strmov(engine_pos, sub_elem->partition_name);
length= (uint)(end - engine_key);
/* Copy the suffix also to query cache key */
memcpy(query_cache_key_end, engine_key_end, (end - engine_key_end));
if (reg_query_cache_dependant_table(thd, engine_key, length,
query_cache_key,
length + diff_length,
m_file[part]->table_cache_type(),
cache,
block_table, m_file[part],
n))
DBUG_RETURN(TRUE);
} while (++j < num_subparts);
}
else
{
char *end= engine_pos+1; // copy end \0
uint length= (uint)(end - engine_key);
/* Copy the suffix also to query cache key */
memcpy(query_cache_key_end, engine_key_end, (end - engine_key_end));
if (reg_query_cache_dependant_table(thd, engine_key, length,
query_cache_key,
length + diff_length,
m_file[i]->table_cache_type(),
cache,
block_table, m_file[i],
n))
DBUG_RETURN(TRUE);
}
} while (++i < num_parts);
DBUG_PRINT("info", ("cnt: %u", (uint)m_tot_parts));
DBUG_RETURN(FALSE);
}
/**
Set up table share object before calling create on underlying handler
@param table Table object
@param info Create info
@param part_elem[in,out] Pointer to used partition_element, searched if NULL
@return status
@retval TRUE Error
@retval FALSE Success
@details
Set up
1) Comment on partition
2) MAX_ROWS, MIN_ROWS on partition
3) Index file name on partition
4) Data file name on partition
*/
int ha_partition::set_up_table_before_create(TABLE *tbl,
const char *partition_name_with_path,
HA_CREATE_INFO *info,
partition_element *part_elem)
{
int error= 0;
const char *partition_name;
THD *thd= ha_thd();
DBUG_ENTER("set_up_table_before_create");
DBUG_ASSERT(part_elem);
if (!part_elem)
DBUG_RETURN(1);
tbl->s->max_rows= part_elem->part_max_rows;
tbl->s->min_rows= part_elem->part_min_rows;
partition_name= strrchr(partition_name_with_path, FN_LIBCHAR);
if ((part_elem->index_file_name &&
(error= append_file_to_dir(thd,
(const char**)&part_elem->index_file_name,
partition_name+1))) ||
(part_elem->data_file_name &&
(error= append_file_to_dir(thd,
(const char**)&part_elem->data_file_name,
partition_name+1))))
{
DBUG_RETURN(error);
}
info->index_file_name= part_elem->index_file_name;
info->data_file_name= part_elem->data_file_name;
info->connect_string= part_elem->connect_string;
if (info->connect_string.length)
info->used_fields|= HA_CREATE_USED_CONNECTION;
tbl->s->connect_string= part_elem->connect_string;
DBUG_RETURN(0);
}
/*
Add two names together
SYNOPSIS
name_add()
out:dest Destination string
first_name First name
sec_name Second name
RETURN VALUE
>0 Error
0 Success
DESCRIPTION
Routine used to add two names with '_' in between then. Service routine
to create_handler_file
Include the NULL in the count of characters since it is needed as separator
between the partition names.
*/
static uint name_add(char *dest, const char *first_name, const char *sec_name)
{
return (uint) (strxmov(dest, first_name, "#SP#", sec_name, NullS) -dest) + 1;
}
/**
Create the special .par file
@param name Full path of table name
@return Operation status
@retval FALSE Error code
@retval TRUE Success
@note
Method used to create handler file with names of partitions, their
engine types and the number of partitions.
*/
bool ha_partition::create_handler_file(const char *name)
{
partition_element *part_elem, *subpart_elem;
uint i, j, part_name_len, subpart_name_len;
uint tot_partition_words, tot_name_len, num_parts;
uint tot_parts= 0;
uint tot_len_words, tot_len_byte, chksum, tot_name_words;
char *name_buffer_ptr;
uchar *file_buffer, *engine_array;
bool result= TRUE;
char file_name[FN_REFLEN];
char part_name[FN_REFLEN];
char subpart_name[FN_REFLEN];
File file;
List_iterator_fast <partition_element> part_it(m_part_info->partitions);
DBUG_ENTER("create_handler_file");
num_parts= m_part_info->partitions.elements;
DBUG_PRINT("info", ("table name = %s, num_parts = %u", name,
num_parts));
tot_name_len= 0;
for (i= 0; i < num_parts; i++)
{
part_elem= part_it++;
if (part_elem->part_state != PART_NORMAL &&
part_elem->part_state != PART_TO_BE_ADDED &&
part_elem->part_state != PART_CHANGED)
continue;
tablename_to_filename(part_elem->partition_name, part_name,
FN_REFLEN);
part_name_len= strlen(part_name);
if (!m_is_sub_partitioned)
{
tot_name_len+= part_name_len + 1;
tot_parts++;
}
else
{
List_iterator_fast <partition_element> sub_it(part_elem->subpartitions);
for (j= 0; j < m_part_info->num_subparts; j++)
{
subpart_elem= sub_it++;
tablename_to_filename(subpart_elem->partition_name,
subpart_name,
FN_REFLEN);
subpart_name_len= strlen(subpart_name);
tot_name_len+= part_name_len + subpart_name_len + 5;
tot_parts++;
}
}
}
/*
File format:
Length in words 4 byte
Checksum 4 byte
Total number of partitions 4 byte
Array of engine types n * 4 bytes where
n = (m_tot_parts + 3)/4
Length of name part in bytes 4 bytes
(Names in filename format)
Name part m * 4 bytes where
m = ((length_name_part + 3)/4)*4
All padding bytes are zeroed
*/
tot_partition_words= (tot_parts + PAR_WORD_SIZE - 1) / PAR_WORD_SIZE;
tot_name_words= (tot_name_len + PAR_WORD_SIZE - 1) / PAR_WORD_SIZE;
/* 4 static words (tot words, checksum, tot partitions, name length) */
tot_len_words= 4 + tot_partition_words + tot_name_words;
tot_len_byte= PAR_WORD_SIZE * tot_len_words;
if (!(file_buffer= (uchar *) my_malloc(tot_len_byte, MYF(MY_ZEROFILL))))
DBUG_RETURN(TRUE);
engine_array= (file_buffer + PAR_ENGINES_OFFSET);
name_buffer_ptr= (char*) (engine_array + tot_partition_words * PAR_WORD_SIZE
+ PAR_WORD_SIZE);
part_it.rewind();
for (i= 0; i < num_parts; i++)
{
part_elem= part_it++;
if (part_elem->part_state != PART_NORMAL &&
part_elem->part_state != PART_TO_BE_ADDED &&
part_elem->part_state != PART_CHANGED)
continue;
if (!m_is_sub_partitioned)
{
tablename_to_filename(part_elem->partition_name, part_name, FN_REFLEN);
name_buffer_ptr= strmov(name_buffer_ptr, part_name)+1;
*engine_array= (uchar) ha_legacy_type(part_elem->engine_type);
DBUG_PRINT("info", ("engine: %u", *engine_array));
engine_array++;
}
else
{
List_iterator_fast <partition_element> sub_it(part_elem->subpartitions);
for (j= 0; j < m_part_info->num_subparts; j++)
{
subpart_elem= sub_it++;
tablename_to_filename(part_elem->partition_name, part_name,
FN_REFLEN);
tablename_to_filename(subpart_elem->partition_name, subpart_name,
FN_REFLEN);
name_buffer_ptr+= name_add(name_buffer_ptr,
part_name,
subpart_name);
*engine_array= (uchar) ha_legacy_type(subpart_elem->engine_type);
DBUG_PRINT("info", ("engine: %u", *engine_array));
engine_array++;
}
}
}
chksum= 0;
int4store(file_buffer, tot_len_words);
int4store(file_buffer + PAR_NUM_PARTS_OFFSET, tot_parts);
int4store(file_buffer + PAR_ENGINES_OFFSET +
(tot_partition_words * PAR_WORD_SIZE),
tot_name_len);
for (i= 0; i < tot_len_words; i++)
chksum^= uint4korr(file_buffer + PAR_WORD_SIZE * i);
int4store(file_buffer + PAR_CHECKSUM_OFFSET, chksum);
/*
Add .par extension to the file name.
Create and write and close file
to be used at open, delete_table and rename_table
*/
fn_format(file_name, name, "", ha_par_ext, MY_APPEND_EXT);
if ((file= mysql_file_create(key_file_partition,
file_name, CREATE_MODE, O_RDWR | O_TRUNC,
MYF(MY_WME))) >= 0)
{
result= mysql_file_write(file, (uchar *) file_buffer, tot_len_byte,
MYF(MY_WME | MY_NABP)) != 0;
/* Write connection information (for federatedx engine) */
part_it.rewind();
for (i= 0; i < num_parts && !result; i++)
{
uchar buffer[4];
part_elem= part_it++;
uint length = part_elem->connect_string.length;
int4store(buffer, length);
if (my_write(file, buffer, 4, MYF(MY_WME | MY_NABP)) ||
my_write(file, (uchar *) part_elem->connect_string.str, length,
MYF(MY_WME | MY_NABP)))
{
result= TRUE;
break;
}
}
(void) mysql_file_close(file, MYF(0));
}
else
result= TRUE;
my_free(file_buffer);
DBUG_RETURN(result);
}
/**
Clear handler variables and free some memory
*/
void ha_partition::clear_handler_file()
{
if (m_engine_array)
plugin_unlock_list(NULL, m_engine_array, m_tot_parts);
free_root(&m_mem_root, MYF(MY_KEEP_PREALLOC));
m_file_buffer= NULL;
m_engine_array= NULL;
m_connect_string= NULL;
}
/**
Create underlying handler objects
@param mem_root Allocate memory through this
@return Operation status
@retval TRUE Error
@retval FALSE Success
*/
bool ha_partition::create_handlers(MEM_ROOT *mem_root)
{
uint i;
uint alloc_len= (m_tot_parts + 1) * sizeof(handler*);
handlerton *hton0;
DBUG_ENTER("create_handlers");
if (!(m_file= (handler **) alloc_root(mem_root, alloc_len)))
DBUG_RETURN(TRUE);
m_file_tot_parts= m_tot_parts;
bzero((char*) m_file, alloc_len);
for (i= 0; i < m_tot_parts; i++)
{
handlerton *hton= plugin_data(m_engine_array[i], handlerton*);
if (!(m_file[i]= get_new_handler(table_share, mem_root, hton)))
DBUG_RETURN(TRUE);
DBUG_PRINT("info", ("engine_type: %u", hton->db_type));
}
/* For the moment we only support partition over the same table engine */
hton0= plugin_data(m_engine_array[0], handlerton*);
if (hton0 == myisam_hton)
{
DBUG_PRINT("info", ("MyISAM"));
m_myisam= TRUE;
}
/* INNODB may not be compiled in... */
else if (ha_legacy_type(hton0) == DB_TYPE_INNODB)
{
DBUG_PRINT("info", ("InnoDB"));
m_innodb= TRUE;
}
DBUG_RETURN(FALSE);
}
/*
Create underlying handler objects from partition info
SYNOPSIS
new_handlers_from_part_info()
mem_root Allocate memory through this
RETURN VALUE
TRUE Error
FALSE Success
*/
bool ha_partition::new_handlers_from_part_info(MEM_ROOT *mem_root)
{
uint i, j, part_count;
partition_element *part_elem;
uint alloc_len= (m_tot_parts + 1) * sizeof(handler*);
List_iterator_fast <partition_element> part_it(m_part_info->partitions);
DBUG_ENTER("ha_partition::new_handlers_from_part_info");
if (!(m_file= (handler **) alloc_root(mem_root, alloc_len)))
{
mem_alloc_error(alloc_len);
goto error_end;
}
m_file_tot_parts= m_tot_parts;
bzero((char*) m_file, alloc_len);
DBUG_ASSERT(m_part_info->num_parts > 0);
i= 0;
part_count= 0;
/*
Don't know the size of the underlying storage engine, invent a number of
bytes allocated for error message if allocation fails
*/
do
{
part_elem= part_it++;
if (m_is_sub_partitioned)
{
for (j= 0; j < m_part_info->num_subparts; j++)
{
if (!(m_file[part_count++]= get_new_handler(table_share, mem_root,
part_elem->engine_type)))
goto error;
DBUG_PRINT("info", ("engine_type: %u",
(uint) ha_legacy_type(part_elem->engine_type)));
}
}
else
{
if (!(m_file[part_count++]= get_new_handler(table_share, mem_root,
part_elem->engine_type)))
goto error;
DBUG_PRINT("info", ("engine_type: %u",
(uint) ha_legacy_type(part_elem->engine_type)));
}
} while (++i < m_part_info->num_parts);
if (part_elem->engine_type == myisam_hton)
{
DBUG_PRINT("info", ("MyISAM"));
m_myisam= TRUE;
}
DBUG_RETURN(FALSE);
error:
mem_alloc_error(sizeof(handler));
error_end:
DBUG_RETURN(TRUE);
}
/**
Read the .par file to get the partitions engines and names
@param name Name of table file (without extention)
@return Operation status
@retval true Failure
@retval false Success
@note On success, m_file_buffer is allocated and must be
freed by the caller. m_name_buffer_ptr and m_tot_parts is also set.
*/
bool ha_partition::read_par_file(const char *name)
{
char buff[FN_REFLEN];
uchar *tot_name_len_offset;
File file;
uchar *file_buffer;
uint i, len_bytes, len_words, tot_partition_words, tot_name_words, chksum;
DBUG_ENTER("ha_partition::read_par_file");
DBUG_PRINT("enter", ("table name: '%s'", name));
if (m_file_buffer)
DBUG_RETURN(false);
fn_format(buff, name, "", ha_par_ext, MY_APPEND_EXT);
/* Following could be done with mysql_file_stat to read in whole file */
if ((file= mysql_file_open(key_file_partition,
buff, O_RDONLY | O_SHARE, MYF(0))) < 0)
DBUG_RETURN(TRUE);
if (mysql_file_read(file, (uchar *) &buff[0], PAR_WORD_SIZE, MYF(MY_NABP)))
goto err1;
len_words= uint4korr(buff);
len_bytes= PAR_WORD_SIZE * len_words;
if (mysql_file_seek(file, 0, MY_SEEK_SET, MYF(0)) == MY_FILEPOS_ERROR)
goto err1;
if (!(file_buffer= (uchar*) alloc_root(&m_mem_root, len_bytes)))
goto err1;
if (mysql_file_read(file, file_buffer, len_bytes, MYF(MY_NABP)))
goto err2;
chksum= 0;
for (i= 0; i < len_words; i++)
chksum ^= uint4korr((file_buffer) + PAR_WORD_SIZE * i);
if (chksum)
goto err2;
m_tot_parts= uint4korr((file_buffer) + PAR_NUM_PARTS_OFFSET);
DBUG_PRINT("info", ("No of parts = %u", m_tot_parts));
tot_partition_words= (m_tot_parts + PAR_WORD_SIZE - 1) / PAR_WORD_SIZE;
tot_name_len_offset= file_buffer + PAR_ENGINES_OFFSET +
PAR_WORD_SIZE * tot_partition_words;
tot_name_words= (uint4korr(tot_name_len_offset) + PAR_WORD_SIZE - 1) /
PAR_WORD_SIZE;
/*
Verify the total length = tot size word, checksum word, num parts word +
engines array + name length word + name array.
*/
if (len_words != (tot_partition_words + tot_name_words + 4))
goto err2;
m_file_buffer= file_buffer; // Will be freed in clear_handler_file()
m_name_buffer_ptr= (char*) (tot_name_len_offset + PAR_WORD_SIZE);
if (!(m_connect_string= (LEX_CSTRING*)
alloc_root(&m_mem_root, m_tot_parts * sizeof(LEX_CSTRING))))
goto err2;
bzero(m_connect_string, m_tot_parts * sizeof(LEX_CSTRING));
/* Read connection arguments (for federated X engine) */
for (i= 0; i < m_tot_parts; i++)
{
LEX_CSTRING connect_string;
uchar buffer[4];
char *tmp;
if (my_read(file, buffer, 4, MYF(MY_NABP)))
{
/* No extra options; Probably not a federatedx engine */
break;
}
connect_string.length= uint4korr(buffer);
connect_string.str= tmp= (char*) alloc_root(&m_mem_root,
connect_string.length+1);
if (my_read(file, (uchar*) connect_string.str, connect_string.length,
MYF(MY_NABP)))
break;
tmp[connect_string.length]= 0;
m_connect_string[i]= connect_string;
}
(void) mysql_file_close(file, MYF(0));
DBUG_RETURN(false);
err2:
err1:
(void) mysql_file_close(file, MYF(0));
DBUG_RETURN(true);
}
/**
Setup m_engine_array
@param mem_root MEM_ROOT to use for allocating new handlers
@return Operation status
@retval false Success
@retval true Failure
*/
bool ha_partition::setup_engine_array(MEM_ROOT *mem_root)
{
uint i;
uchar *buff;
handlerton **engine_array, *first_engine;
enum legacy_db_type db_type, first_db_type;
DBUG_ASSERT(!m_file);
DBUG_ENTER("ha_partition::setup_engine_array");
engine_array= (handlerton **) my_alloca(m_tot_parts * sizeof(handlerton*));
if (!engine_array)
DBUG_RETURN(true);
buff= (uchar *) (m_file_buffer + PAR_ENGINES_OFFSET);
first_db_type= (enum legacy_db_type) buff[0];
first_engine= ha_resolve_by_legacy_type(ha_thd(), first_db_type);
if (!first_engine)
goto err;
if (!(m_engine_array= (plugin_ref*)
alloc_root(&m_mem_root, m_tot_parts * sizeof(plugin_ref))))
goto err;
for (i= 0; i < m_tot_parts; i++)
{
db_type= (enum legacy_db_type) buff[i];
if (db_type != first_db_type)
{
DBUG_PRINT("error", ("partition %u engine %d is not same as "
"first partition %d", i, db_type,
(int) first_db_type));
DBUG_ASSERT(0);
clear_handler_file();
goto err;
}
m_engine_array[i]= ha_lock_engine(NULL, first_engine);
if (!m_engine_array[i])
{
clear_handler_file();
goto err;
}
}
my_afree(engine_array);
if (create_handlers(mem_root))
{
clear_handler_file();
DBUG_RETURN(true);
}
DBUG_RETURN(false);
err:
my_afree(engine_array);
DBUG_RETURN(true);
}
/**
Get info about partition engines and their names from the .par file
@param name Full path of table name
@param mem_root Allocate memory through this
@param is_clone If it is a clone, don't create new handlers
@return Operation status
@retval true Error
@retval false Success
@note Open handler file to get partition names, engine types and number of
partitions.
*/
bool ha_partition::get_from_handler_file(const char *name, MEM_ROOT *mem_root,
bool is_clone)
{
DBUG_ENTER("ha_partition::get_from_handler_file");
DBUG_PRINT("enter", ("table name: '%s'", name));
if (m_file_buffer)
DBUG_RETURN(false);
if (read_par_file(name))
DBUG_RETURN(true);
if (!is_clone && setup_engine_array(mem_root))
DBUG_RETURN(true);
DBUG_RETURN(false);
}
/****************************************************************************
MODULE open/close object
****************************************************************************/
/**
Get the partition name.
@param part Struct containing name and length
@param[out] length Length of the name
@return Partition name
*/
static uchar *get_part_name(PART_NAME_DEF *part, size_t *length,
my_bool not_used __attribute__((unused)))
{
*length= part->length;
return part->partition_name;
}
/**
Insert a partition name in the partition_name_hash.
@param name Name of partition
@param part_id Partition id (number)
@param is_subpart Set if the name belongs to a subpartition
@return Operation status
@retval true Failure
@retval false Sucess
*/
bool ha_partition::insert_partition_name_in_hash(const char *name, uint part_id,
bool is_subpart)
{
PART_NAME_DEF *part_def;
uchar *part_name;
uint part_name_length;
DBUG_ENTER("ha_partition::insert_partition_name_in_hash");
/*
Calculate and store the length here, to avoid doing it when
searching the hash.
*/
part_name_length= strlen(name);
/*
Must use memory that lives as long as table_share.
Freed in the Partition_share destructor.
Since we use my_multi_malloc, then my_free(part_def) will also free
part_name, as a part of my_hash_free.
*/
if (!my_multi_malloc(MY_WME,
&part_def, sizeof(PART_NAME_DEF),
&part_name, part_name_length + 1,
NULL))
DBUG_RETURN(true);
memcpy(part_name, name, part_name_length + 1);
part_def->partition_name= part_name;
part_def->length= part_name_length;
part_def->part_id= part_id;
part_def->is_subpart= is_subpart;
if (my_hash_insert(&part_share->partition_name_hash, (uchar *) part_def))
{
my_free(part_def);
DBUG_RETURN(true);
}
DBUG_RETURN(false);
}
/**
Populate the partition_name_hash in part_share.
*/
bool ha_partition::populate_partition_name_hash()
{
List_iterator<partition_element> part_it(m_part_info->partitions);
uint num_parts= m_part_info->num_parts;
uint num_subparts= m_is_sub_partitioned ? m_part_info->num_subparts : 1;
uint tot_names;
uint i= 0;
DBUG_ASSERT(part_share);
DBUG_ENTER("ha_partition::populate_partition_name_hash");
/*
partition_name_hash is only set once and never changed
-> OK to check without locking.
*/
if (part_share->partition_name_hash_initialized)
DBUG_RETURN(false);
lock_shared_ha_data();
if (part_share->partition_name_hash_initialized)
{
unlock_shared_ha_data();
DBUG_RETURN(false);
}
tot_names= m_is_sub_partitioned ? m_tot_parts + num_parts : num_parts;
if (my_hash_init(&part_share->partition_name_hash,
system_charset_info, tot_names, 0, 0,
(my_hash_get_key) get_part_name,
my_free, HASH_UNIQUE))
{
unlock_shared_ha_data();
DBUG_RETURN(TRUE);
}
do
{
partition_element *part_elem= part_it++;
DBUG_ASSERT(part_elem->part_state == PART_NORMAL);
if (part_elem->part_state == PART_NORMAL)
{
if (insert_partition_name_in_hash(part_elem->partition_name,
i * num_subparts, false))
goto err;
if (m_is_sub_partitioned)
{
List_iterator<partition_element>
subpart_it(part_elem->subpartitions);
partition_element *sub_elem;
uint j= 0;
do
{
sub_elem= subpart_it++;
if (insert_partition_name_in_hash(sub_elem->partition_name,
i * num_subparts + j, true))
goto err;
} while (++j < num_subparts);
}
}
} while (++i < num_parts);
part_share->partition_name_hash_initialized= true;
unlock_shared_ha_data();
DBUG_RETURN(FALSE);
err:
my_hash_free(&part_share->partition_name_hash);
unlock_shared_ha_data();
DBUG_RETURN(TRUE);
}
/**
Set Handler_share pointer and allocate Handler_share pointers
for each partition and set those.
@param ha_share_arg Where to store/retrieve the Partitioning_share pointer
to be shared by all instances of the same table.
@return Operation status
@retval true Failure
@retval false Sucess
*/
bool ha_partition::set_ha_share_ref(Handler_share **ha_share_arg)
{
Handler_share **ha_shares;
uint i;
DBUG_ENTER("ha_partition::set_ha_share_ref");
DBUG_ASSERT(!part_share);
DBUG_ASSERT(table_share);
DBUG_ASSERT(!m_is_clone_of);
DBUG_ASSERT(m_tot_parts);
if (handler::set_ha_share_ref(ha_share_arg))
DBUG_RETURN(true);
if (!(part_share= get_share()))
DBUG_RETURN(true);
DBUG_ASSERT(part_share->partitions_share_refs.num_parts >= m_tot_parts);
ha_shares= part_share->partitions_share_refs.ha_shares;
for (i= 0; i < m_tot_parts; i++)
{
if (m_file[i]->set_ha_share_ref(&ha_shares[i]))
DBUG_RETURN(true);
}
DBUG_RETURN(false);
}
/**
Get the PARTITION_SHARE for the table.
@return Operation status
@retval true Error
@retval false Success
@note Gets or initializes the Partition_share object used by partitioning.
The Partition_share is used for handling the auto_increment etc.
*/
Partition_share *ha_partition::get_share()
{
Partition_share *tmp_share;
DBUG_ENTER("ha_partition::get_share");
DBUG_ASSERT(table_share);
lock_shared_ha_data();
if (!(tmp_share= static_cast<Partition_share*>(get_ha_share_ptr())))
{
tmp_share= new Partition_share;
if (!tmp_share)
goto err;
if (tmp_share->init(m_tot_parts))
{
delete tmp_share;
tmp_share= NULL;
goto err;
}
set_ha_share_ptr(static_cast<Handler_share*>(tmp_share));
}
err:
unlock_shared_ha_data();
DBUG_RETURN(tmp_share);
}
/**
Helper function for freeing all internal bitmaps.
*/
void ha_partition::free_partition_bitmaps()
{
/* Initialize the bitmap we use to minimize ha_start_bulk_insert calls */
my_bitmap_free(&m_bulk_insert_started);
my_bitmap_free(&m_locked_partitions);
my_bitmap_free(&m_partitions_to_reset);
my_bitmap_free(&m_key_not_found_partitions);
}
/**
Helper function for initializing all internal bitmaps.
*/
bool ha_partition::init_partition_bitmaps()
{
DBUG_ENTER("ha_partition::init_partition_bitmaps");
/* Initialize the bitmap we use to minimize ha_start_bulk_insert calls */
if (my_bitmap_init(&m_bulk_insert_started, NULL, m_tot_parts + 1, FALSE))
DBUG_RETURN(true);
bitmap_clear_all(&m_bulk_insert_started);
/* Initialize the bitmap we use to keep track of locked partitions */
if (my_bitmap_init(&m_locked_partitions, NULL, m_tot_parts, FALSE))
{
my_bitmap_free(&m_bulk_insert_started);
DBUG_RETURN(true);
}
bitmap_clear_all(&m_locked_partitions);
/*
Initialize the bitmap we use to keep track of partitions which may have
something to reset in ha_reset().
*/
if (my_bitmap_init(&m_partitions_to_reset, NULL, m_tot_parts, FALSE))
{
my_bitmap_free(&m_bulk_insert_started);
my_bitmap_free(&m_locked_partitions);
DBUG_RETURN(true);
}
bitmap_clear_all(&m_partitions_to_reset);
/*
Initialize the bitmap we use to keep track of partitions which returned
HA_ERR_KEY_NOT_FOUND from index_read_map.
*/
if (my_bitmap_init(&m_key_not_found_partitions, NULL, m_tot_parts, FALSE))
{
my_bitmap_free(&m_bulk_insert_started);
my_bitmap_free(&m_locked_partitions);
my_bitmap_free(&m_partitions_to_reset);
DBUG_RETURN(true);
}
bitmap_clear_all(&m_key_not_found_partitions);
m_key_not_found= false;
/* Initialize the bitmap for read/lock_partitions */
if (!m_is_clone_of)
{
DBUG_ASSERT(!m_clone_mem_root);
if (m_part_info->set_partition_bitmaps(NULL))
{
free_partition_bitmaps();
DBUG_RETURN(true);
}
}
DBUG_RETURN(false);
}
/*
Open handler object
SYNOPSIS
open()
name Full path of table name
mode Open mode flags
test_if_locked ?
RETURN VALUE
>0 Error
0 Success
DESCRIPTION
Used for opening tables. The name will be the name of the file.
A table is opened when it needs to be opened. For instance
when a request comes in for a select on the table (tables are not
open and closed for each request, they are cached).
Called from handler.cc by handler::ha_open(). The server opens all tables
by calling ha_open() which then calls the handler specific open().
*/
int ha_partition::open(const char *name, int mode, uint test_if_locked)
{
char *name_buffer_ptr;
int error= HA_ERR_INITIALIZATION;
handler **file;
char name_buff[FN_REFLEN + 1];
ulonglong check_table_flags;
DBUG_ENTER("ha_partition::open");
DBUG_ASSERT(table->s == table_share);
ref_length= 0;
m_mode= mode;
m_open_test_lock= test_if_locked;
m_part_field_array= m_part_info->full_part_field_array;
if (get_from_handler_file(name, &table->mem_root, MY_TEST(m_is_clone_of)))
DBUG_RETURN(error);
name_buffer_ptr= m_name_buffer_ptr;
if (populate_partition_name_hash())
{
DBUG_RETURN(HA_ERR_INITIALIZATION);
}
m_start_key.length= 0;
m_rec0= table->record[0];
m_rec_length= table_share->reclength;
if (!m_part_ids_sorted_by_num_of_records)
{
if (!(m_part_ids_sorted_by_num_of_records=
(uint32*) my_malloc(m_tot_parts * sizeof(uint32), MYF(MY_WME))))
DBUG_RETURN(error);
uint32 i;
/* Initialize it with all partition ids. */
for (i= 0; i < m_tot_parts; i++)
m_part_ids_sorted_by_num_of_records[i]= i;
}
if (init_partition_bitmaps())
DBUG_RETURN(error);
DBUG_ASSERT(m_part_info);
if (m_is_clone_of)
{
uint i, alloc_len;
DBUG_ASSERT(m_clone_mem_root);
/* Allocate an array of handler pointers for the partitions handlers. */
alloc_len= (m_tot_parts + 1) * sizeof(handler*);
if (!(m_file= (handler **) alloc_root(m_clone_mem_root, alloc_len)))
{
error= HA_ERR_INITIALIZATION;
goto err_alloc;
}
memset(m_file, 0, alloc_len);
/*
Populate them by cloning the original partitions. This also opens them.
Note that file->ref is allocated too.
*/
file= m_is_clone_of->m_file;
for (i= 0; i < m_tot_parts; i++)
{
if ((error= create_partition_name(name_buff, sizeof(name_buff), name,
name_buffer_ptr, NORMAL_PART_NAME, FALSE)))
goto err_handler;
/* ::clone() will also set ha_share from the original. */
if (!(m_file[i]= file[i]->clone(name_buff, m_clone_mem_root)))
{
error= HA_ERR_INITIALIZATION;
file= &m_file[i];
goto err_handler;
}
name_buffer_ptr+= strlen(name_buffer_ptr) + 1;
}
}
else
{
file= m_file;
do
{
if ((error= create_partition_name(name_buff, sizeof(name_buff), name,
name_buffer_ptr, NORMAL_PART_NAME, FALSE)))
goto err_handler;
table->s->connect_string = m_connect_string[(uint)(file-m_file)];
if ((error= (*file)->ha_open(table, name_buff, mode,
test_if_locked | HA_OPEN_NO_PSI_CALL)))
goto err_handler;
bzero(&table->s->connect_string, sizeof(LEX_STRING));
if (m_file == file)
m_num_locks= (*file)->lock_count();
DBUG_ASSERT(m_num_locks == (*file)->lock_count());
name_buffer_ptr+= strlen(name_buffer_ptr) + 1;
} while (*(++file));
}
file= m_file;
ref_length= (*file)->ref_length;
check_table_flags= (((*file)->ha_table_flags() &
~(PARTITION_DISABLED_TABLE_FLAGS)) |
(PARTITION_ENABLED_TABLE_FLAGS));
while (*(++file))
{
/* MyISAM can have smaller ref_length for partitions with MAX_ROWS set */
set_if_bigger(ref_length, ((*file)->ref_length));
/*
Verify that all partitions have the same set of table flags.
Mask all flags that partitioning enables/disables.
*/
if (check_table_flags != (((*file)->ha_table_flags() &
~(PARTITION_DISABLED_TABLE_FLAGS)) |
(PARTITION_ENABLED_TABLE_FLAGS)))
{
error= HA_ERR_INITIALIZATION;
/* set file to last handler, so all of them are closed */
file = &m_file[m_tot_parts - 1];
goto err_handler;
}
}
key_used_on_scan= m_file[0]->key_used_on_scan;
implicit_emptied= m_file[0]->implicit_emptied;
/*
Add 2 bytes for partition id in position ref length.
ref_length=max_in_all_partitions(ref_length) + PARTITION_BYTES_IN_POS
*/
ref_length+= PARTITION_BYTES_IN_POS;
m_ref_length= ref_length;
/*
Release buffer read from .par file. It will not be reused again after
being opened once.
*/
clear_handler_file();
/*
Some handlers update statistics as part of the open call. This will in
some cases corrupt the statistics of the partition handler and thus
to ensure we have correct statistics we call info from open after
calling open on all individual handlers.
*/
m_handler_status= handler_opened;
if (m_part_info->part_expr)
m_part_func_monotonicity_info=
m_part_info->part_expr->get_monotonicity_info();
else if (m_part_info->list_of_part_fields)
m_part_func_monotonicity_info= MONOTONIC_STRICT_INCREASING;
info(HA_STATUS_VARIABLE | HA_STATUS_CONST);
DBUG_RETURN(0);
err_handler:
DEBUG_SYNC(ha_thd(), "partition_open_error");
while (file-- != m_file)
(*file)->ha_close();
err_alloc:
free_partition_bitmaps();
DBUG_RETURN(error);
}
/*
Disabled since it is not possible to prune yet.
without pruning, it need to rebind/unbind every partition in every
statement which uses a table from the table cache. Will also use
as many PSI_tables as there are partitions.
*/
#ifdef HAVE_M_PSI_PER_PARTITION
void ha_partition::unbind_psi()
{
uint i;
DBUG_ENTER("ha_partition::unbind_psi");
handler::unbind_psi();
for (i= 0; i < m_tot_parts; i++)
{
DBUG_ASSERT(m_file[i] != NULL);
m_file[i]->unbind_psi();
}
DBUG_VOID_RETURN;
}
void ha_partition::rebind_psi()
{
uint i;
DBUG_ENTER("ha_partition::rebind_psi");
handler::rebind_psi();
for (i= 0; i < m_tot_parts; i++)
{
DBUG_ASSERT(m_file[i] != NULL);
m_file[i]->rebind_psi();
}
DBUG_VOID_RETURN;
}
#endif /* HAVE_M_PSI_PER_PARTITION */
/**
Clone the open and locked partitioning handler.
@param mem_root MEM_ROOT to use.
@return Pointer to the successfully created clone or NULL
@details
This function creates a new ha_partition handler as a clone/copy. The
original (this) must already be opened and locked. The clone will use
the originals m_part_info.
It also allocates memory for ref + ref_dup.
In ha_partition::open() it will clone its original handlers partitions
which will allocate then on the correct MEM_ROOT and also open them.
*/
handler *ha_partition::clone(const char *name, MEM_ROOT *mem_root)
{
ha_partition *new_handler;
DBUG_ENTER("ha_partition::clone");
new_handler= new (mem_root) ha_partition(ht, table_share, m_part_info,
this, mem_root);
if (!new_handler)
DBUG_RETURN(NULL);
/*
We will not clone each partition's handler here, it will be done in
ha_partition::open() for clones. Also set_ha_share_ref is not needed
here, since 1) ha_share is copied in the constructor used above
2) each partition's cloned handler will set it from its original.
*/
/*
Allocate new_handler->ref here because otherwise ha_open will allocate it
on this->table->mem_root and we will not be able to reclaim that memory
when the clone handler object is destroyed.
*/
if (!(new_handler->ref= (uchar*) alloc_root(mem_root,
ALIGN_SIZE(m_ref_length)*2)))
goto err;
if (new_handler->ha_open(table, name,
table->db_stat,
HA_OPEN_IGNORE_IF_LOCKED | HA_OPEN_NO_PSI_CALL))
goto err;
DBUG_RETURN((handler*) new_handler);
err:
delete new_handler;
DBUG_RETURN(NULL);
}
/*
Close handler object
SYNOPSIS
close()
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
Called from sql_base.cc, sql_select.cc, and table.cc.
In sql_select.cc it is only used to close up temporary tables or during
the process where a temporary table is converted over to being a
myisam table.
For sql_base.cc look at close_data_tables().
*/
int ha_partition::close(void)
{
bool first= TRUE;
handler **file;
DBUG_ENTER("ha_partition::close");
DBUG_ASSERT(table->s == table_share);
destroy_record_priority_queue();
free_partition_bitmaps();
DBUG_ASSERT(m_part_info);
file= m_file;
repeat:
do
{
(*file)->ha_close();
} while (*(++file));
if (first && m_added_file && m_added_file[0])
{
file= m_added_file;
first= FALSE;
goto repeat;
}
m_handler_status= handler_closed;
DBUG_RETURN(0);
}
/****************************************************************************
MODULE start/end statement
****************************************************************************/
/*
A number of methods to define various constants for the handler. In
the case of the partition handler we need to use some max and min
of the underlying handlers in most cases.
*/
/*
Set external locks on table
SYNOPSIS
external_lock()
thd Thread object
lock_type Type of external lock
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
First you should go read the section "locking functions for mysql" in
lock.cc to understand this.
This create a lock on the table. If you are implementing a storage engine
that can handle transactions look at ha_berkeley.cc to see how you will
want to go about doing this. Otherwise you should consider calling
flock() here.
Originally this method was used to set locks on file level to enable
several MySQL Servers to work on the same data. For transactional
engines it has been "abused" to also mean start and end of statements
to enable proper rollback of statements and transactions. When LOCK
TABLES has been issued the start_stmt method takes over the role of
indicating start of statement but in this case there is no end of
statement indicator(?).
Called from lock.cc by lock_external() and unlock_external(). Also called
from sql_table.cc by copy_data_between_tables().
*/
int ha_partition::external_lock(THD *thd, int lock_type)
{
uint error;
uint i, first_used_partition;
MY_BITMAP *used_partitions;
DBUG_ENTER("ha_partition::external_lock");
DBUG_ASSERT(!auto_increment_lock && !auto_increment_safe_stmt_log_lock);
if (lock_type == F_UNLCK)
used_partitions= &m_locked_partitions;
else
used_partitions= &(m_part_info->lock_partitions);
first_used_partition= bitmap_get_first_set(used_partitions);
for (i= first_used_partition;
i < m_tot_parts;
i= bitmap_get_next_set(used_partitions, i))
{
DBUG_PRINT("info", ("external_lock(thd, %d) part %d", lock_type, i));
if ((error= m_file[i]->ha_external_lock(thd, lock_type)))
{
if (lock_type != F_UNLCK)
goto err_handler;
}
DBUG_PRINT("info", ("external_lock part %u lock %d", i, lock_type));
if (lock_type != F_UNLCK)
bitmap_set_bit(&m_locked_partitions, i);
}
if (lock_type == F_UNLCK)
{
bitmap_clear_all(used_partitions);
}
else
{
/* Add touched partitions to be included in reset(). */
bitmap_union(&m_partitions_to_reset, used_partitions);
}
if (m_added_file && m_added_file[0])
{
handler **file= m_added_file;
DBUG_ASSERT(lock_type == F_UNLCK);
do
{
(void) (*file)->ha_external_lock(thd, lock_type);
} while (*(++file));
}
if (lock_type == F_WRLCK && m_part_info->part_expr)
m_part_info->part_expr->walk(&Item::register_field_in_read_map, 1, 0);
DBUG_RETURN(0);
err_handler:
uint j;
for (j= first_used_partition;
j < i;
j= bitmap_get_next_set(&m_locked_partitions, j))
{
(void) m_file[j]->ha_external_lock(thd, F_UNLCK);
}
bitmap_clear_all(&m_locked_partitions);
DBUG_RETURN(error);
}
/*
Get the lock(s) for the table and perform conversion of locks if needed
SYNOPSIS
store_lock()
thd Thread object
to Lock object array
lock_type Table lock type
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
The idea with handler::store_lock() is the following:
The statement decided which locks we should need for the table
for updates/deletes/inserts we get WRITE locks, for SELECT... we get
read locks.
Before adding the lock into the table lock handler (see thr_lock.c)
mysqld calls store lock with the requested locks. Store lock can now
modify a write lock to a read lock (or some other lock), ignore the
lock (if we don't want to use MySQL table locks at all) or add locks
for many tables (like we do when we are using a MERGE handler).
Berkeley DB for partition changes all WRITE locks to TL_WRITE_ALLOW_WRITE
(which signals that we are doing WRITES, but we are still allowing other
reader's and writer's.
When releasing locks, store_lock() is also called. In this case one
usually doesn't have to do anything.
store_lock is called when holding a global mutex to ensure that only
one thread at a time changes the locking information of tables.
In some exceptional cases MySQL may send a request for a TL_IGNORE;
This means that we are requesting the same lock as last time and this
should also be ignored. (This may happen when someone does a flush
table when we have opened a part of the tables, in which case mysqld
closes and reopens the tables and tries to get the same locks as last
time). In the future we will probably try to remove this.
Called from lock.cc by get_lock_data().
*/
THR_LOCK_DATA **ha_partition::store_lock(THD *thd,
THR_LOCK_DATA **to,
enum thr_lock_type lock_type)
{
uint i;
DBUG_ENTER("ha_partition::store_lock");
DBUG_ASSERT(thd == current_thd);
/*
This can be called from get_lock_data() in mysql_lock_abort_for_thread(),
even when thd != table->in_use. In that case don't use partition pruning,
but use all partitions instead to avoid using another threads structures.
*/
if (thd != table->in_use)
{
for (i= 0; i < m_tot_parts; i++)
to= m_file[i]->store_lock(thd, to, lock_type);
}
else
{
for (i= bitmap_get_first_set(&(m_part_info->lock_partitions));
i < m_tot_parts;
i= bitmap_get_next_set(&m_part_info->lock_partitions, i))
{
DBUG_PRINT("info", ("store lock %d iteration", i));
to= m_file[i]->store_lock(thd, to, lock_type);
}
}
DBUG_RETURN(to);
}
/*
Start a statement when table is locked
SYNOPSIS
start_stmt()
thd Thread object
lock_type Type of external lock
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
This method is called instead of external lock when the table is locked
before the statement is executed.
*/
int ha_partition::start_stmt(THD *thd, thr_lock_type lock_type)
{
int error= 0;
uint i;
/* Assert that read_partitions is included in lock_partitions */
DBUG_ASSERT(bitmap_is_subset(&m_part_info->read_partitions,
&m_part_info->lock_partitions));
/*
m_locked_partitions is set in previous external_lock/LOCK TABLES.
Current statement's lock requests must not include any partitions
not previously locked.
*/
DBUG_ASSERT(bitmap_is_subset(&m_part_info->lock_partitions,
&m_locked_partitions));
DBUG_ENTER("ha_partition::start_stmt");
for (i= bitmap_get_first_set(&(m_part_info->lock_partitions));
i < m_tot_parts;
i= bitmap_get_next_set(&m_part_info->lock_partitions, i))
{
if ((error= m_file[i]->start_stmt(thd, lock_type)))
break;
/* Add partition to be called in reset(). */
bitmap_set_bit(&m_partitions_to_reset, i);
}
if (lock_type == F_WRLCK && m_part_info->part_expr)
m_part_info->part_expr->walk(&Item::register_field_in_read_map, 1, 0);
DBUG_RETURN(error);
}
/**
Get number of lock objects returned in store_lock
@returns Number of locks returned in call to store_lock
@desc
Returns the number of store locks needed in call to store lock.
We return number of partitions we will lock multiplied with number of
locks needed by each partition. Assists the above functions in allocating
sufficient space for lock structures.
*/
uint ha_partition::lock_count() const
{
DBUG_ENTER("ha_partition::lock_count");
/*
The caller want to know the upper bound, to allocate enough memory.
There is no performance lost if we simply return maximum number locks
needed, only some minor over allocation of memory in get_lock_data().
Also notice that this may be called for another thread != table->in_use,
when mysql_lock_abort_for_thread() is called. So this is more safe, then
using number of partitions after pruning.
*/
DBUG_RETURN(m_tot_parts * m_num_locks);
}
/*
Unlock last accessed row
SYNOPSIS
unlock_row()
RETURN VALUE
NONE
DESCRIPTION
Record currently processed was not in the result set of the statement
and is thus unlocked. Used for UPDATE and DELETE queries.
*/
void ha_partition::unlock_row()
{
DBUG_ENTER("ha_partition::unlock_row");
m_file[m_last_part]->unlock_row();
DBUG_VOID_RETURN;
}
/**
Check if semi consistent read was used
SYNOPSIS
was_semi_consistent_read()
RETURN VALUE
TRUE Previous read was a semi consistent read
FALSE Previous read was not a semi consistent read
DESCRIPTION
See handler.h:
In an UPDATE or DELETE, if the row under the cursor was locked by another
transaction, and the engine used an optimistic read of the last
committed row value under the cursor, then the engine returns 1 from this
function. MySQL must NOT try to update this optimistic value. If the
optimistic value does not match the WHERE condition, MySQL can decide to
skip over this row. Currently only works for InnoDB. This can be used to
avoid unnecessary lock waits.
If this method returns nonzero, it will also signal the storage
engine that the next read will be a locking re-read of the row.
*/
bool ha_partition::was_semi_consistent_read()
{
DBUG_ENTER("ha_partition::was_semi_consistent_read");
DBUG_ASSERT(m_last_part < m_tot_parts &&
bitmap_is_set(&(m_part_info->read_partitions), m_last_part));
DBUG_RETURN(m_file[m_last_part]->was_semi_consistent_read());
}
/**
Use semi consistent read if possible
SYNOPSIS
try_semi_consistent_read()
yes Turn on semi consistent read
RETURN VALUE
NONE
DESCRIPTION
See handler.h:
Tell the engine whether it should avoid unnecessary lock waits.
If yes, in an UPDATE or DELETE, if the row under the cursor was locked
by another transaction, the engine may try an optimistic read of
the last committed row value under the cursor.
Note: prune_partitions are already called before this call, so using
pruning is OK.
*/
void ha_partition::try_semi_consistent_read(bool yes)
{
uint i;
DBUG_ENTER("ha_partition::try_semi_consistent_read");
i= bitmap_get_first_set(&(m_part_info->read_partitions));
DBUG_ASSERT(i != MY_BIT_NONE);
for (;
i < m_tot_parts;
i= bitmap_get_next_set(&m_part_info->read_partitions, i))
{
m_file[i]->try_semi_consistent_read(yes);
}
DBUG_VOID_RETURN;
}
/****************************************************************************
MODULE change record
****************************************************************************/
/*
Insert a row to the table
SYNOPSIS
write_row()
buf The row in MySQL Row Format
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
write_row() inserts a row. buf() is a byte array of data, normally
record[0].
You can use the field information to extract the data from the native byte
array type.
Example of this would be:
for (Field **field=table->field ; *field ; field++)
{
...
}
See ha_tina.cc for a variant of extracting all of the data as strings.
ha_berkeley.cc has a variant of how to store it intact by "packing" it
for ha_berkeley's own native storage type.
Called from item_sum.cc, item_sum.cc, sql_acl.cc, sql_insert.cc,
sql_insert.cc, sql_select.cc, sql_table.cc, sql_udf.cc, and sql_update.cc.
ADDITIONAL INFO:
We have to set auto_increment fields, because those may be used in
determining which partition the row should be written to.
*/
int ha_partition::write_row(uchar * buf)
{
uint32 part_id;
int error;
longlong func_value;
bool have_auto_increment= table->next_number_field && buf == table->record[0];
my_bitmap_map *old_map;
THD *thd= ha_thd();
sql_mode_t saved_sql_mode= thd->variables.sql_mode;
bool saved_auto_inc_field_not_null= table->auto_increment_field_not_null;
DBUG_ENTER("ha_partition::write_row");
DBUG_ASSERT(buf == m_rec0);
/*
If we have an auto_increment column and we are writing a changed row
or a new row, then update the auto_increment value in the record.
*/
if (have_auto_increment)
{
if (!part_share->auto_inc_initialized &&
!table_share->next_number_keypart)
{
/*
If auto_increment in table_share is not initialized, start by
initializing it.
*/
info(HA_STATUS_AUTO);
}
error= update_auto_increment();
/*
If we have failed to set the auto-increment value for this row,
it is highly likely that we will not be able to insert it into
the correct partition. We must check and fail if neccessary.
*/
if (error)
goto exit;
/*
Don't allow generation of auto_increment value the partitions handler.
If a partitions handler would change the value, then it might not
match the partition any longer.
This can occur if 'SET INSERT_ID = 0; INSERT (NULL)',
So allow this by adding 'MODE_NO_AUTO_VALUE_ON_ZERO' to sql_mode.
The partitions handler::next_insert_id must always be 0. Otherwise
we need to forward release_auto_increment, or reset it for all
partitions.
*/
if (table->next_number_field->val_int() == 0)
{
table->auto_increment_field_not_null= TRUE;
thd->variables.sql_mode|= MODE_NO_AUTO_VALUE_ON_ZERO;
}
}
old_map= dbug_tmp_use_all_columns(table, table->read_set);
error= m_part_info->get_partition_id(m_part_info, &part_id, &func_value);
dbug_tmp_restore_column_map(table->read_set, old_map);
if (unlikely(error))
{
m_part_info->err_value= func_value;
goto exit;
}
if (!bitmap_is_set(&(m_part_info->lock_partitions), part_id))
{
DBUG_PRINT("info", ("Write to non-locked partition %u (func_value: %ld)",
part_id, (long) func_value));
error= HA_ERR_NOT_IN_LOCK_PARTITIONS;
goto exit;
}
m_last_part= part_id;
DBUG_PRINT("info", ("Insert in partition %d", part_id));
start_part_bulk_insert(thd, part_id);
tmp_disable_binlog(thd); /* Do not replicate the low-level changes. */
error= m_file[part_id]->ha_write_row(buf);
if (have_auto_increment && !table->s->next_number_keypart)
set_auto_increment_if_higher(table->next_number_field);
reenable_binlog(thd);
exit:
thd->variables.sql_mode= saved_sql_mode;
table->auto_increment_field_not_null= saved_auto_inc_field_not_null;
DBUG_RETURN(error);
}
/*
Update an existing row
SYNOPSIS
update_row()
old_data Old record in MySQL Row Format
new_data New record in MySQL Row Format
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
Yes, update_row() does what you expect, it updates a row. old_data will
have the previous row record in it, while new_data will have the newest
data in it.
Keep in mind that the server can do updates based on ordering if an
ORDER BY clause was used. Consecutive ordering is not guarenteed.
Called from sql_select.cc, sql_acl.cc, sql_update.cc, and sql_insert.cc.
new_data is always record[0]
old_data is always record[1]
*/
int ha_partition::update_row(const uchar *old_data, const uchar *new_data)
{
THD *thd= ha_thd();
uint32 new_part_id, old_part_id;
int error= 0;
longlong func_value;
DBUG_ENTER("ha_partition::update_row");
m_err_rec= NULL;
// Need to read partition-related columns, to locate the row's partition:
DBUG_ASSERT(bitmap_is_subset(&m_part_info->full_part_field_set,
table->read_set));
if ((error= get_parts_for_update(old_data, new_data, table->record[0],
m_part_info, &old_part_id, &new_part_id,
&func_value)))
{
m_part_info->err_value= func_value;
goto exit;
}
DBUG_ASSERT(bitmap_is_set(&(m_part_info->read_partitions), old_part_id));
if (!bitmap_is_set(&(m_part_info->lock_partitions), new_part_id))
{
error= HA_ERR_NOT_IN_LOCK_PARTITIONS;
goto exit;
}
/*
The protocol for updating a row is:
1) position the handler (cursor) on the row to be updated,
either through the last read row (rnd or index) or by rnd_pos.
2) call update_row with both old and new full records as arguments.
This means that m_last_part should already be set to actual partition
where the row was read from. And if that is not the same as the
calculated part_id we found a misplaced row, we return an error to
notify the user that something is broken in the row distribution
between partitions! Since we don't check all rows on read, we return an
error instead of correcting m_last_part, to make the user aware of the
problem!
Notice that HA_READ_BEFORE_WRITE_REMOVAL does not require this protocol,
so this is not supported for this engine.
*/
if (old_part_id != m_last_part)
{
m_err_rec= old_data;
DBUG_RETURN(HA_ERR_ROW_IN_WRONG_PARTITION);
}
m_last_part= new_part_id;
start_part_bulk_insert(thd, new_part_id);
if (new_part_id == old_part_id)
{
DBUG_PRINT("info", ("Update in partition %d", new_part_id));
tmp_disable_binlog(thd); /* Do not replicate the low-level changes. */
error= m_file[new_part_id]->ha_update_row(old_data, new_data);
reenable_binlog(thd);
goto exit;
}
else
{
Field *saved_next_number_field= table->next_number_field;
/*
Don't allow generation of auto_increment value for update.
table->next_number_field is never set on UPDATE.
But is set for INSERT ... ON DUPLICATE KEY UPDATE,
and since update_row() does not generate or update an auto_inc value,
we cannot have next_number_field set when moving a row
to another partition with write_row(), since that could
generate/update the auto_inc value.
This gives the same behavior for partitioned vs non partitioned tables.
*/
table->next_number_field= NULL;
DBUG_PRINT("info", ("Update from partition %d to partition %d",
old_part_id, new_part_id));
tmp_disable_binlog(thd); /* Do not replicate the low-level changes. */
error= m_file[new_part_id]->ha_write_row((uchar*) new_data);
reenable_binlog(thd);
table->next_number_field= saved_next_number_field;
if (error)
goto exit;
if (m_part_info->part_type == VERSIONING_PARTITION)
{
uint sub_factor= m_part_info->num_subparts ? m_part_info->num_subparts : 1;
DBUG_ASSERT(m_tot_parts == m_part_info->num_parts * sub_factor);
uint lpart_id= new_part_id / sub_factor;
// lpart_id is VERSIONING partition because new_part_id != old_part_id
m_part_info->vers_update_stats(thd, lpart_id);
}
tmp_disable_binlog(thd); /* Do not replicate the low-level changes. */
error= m_file[old_part_id]->ha_delete_row(old_data);
reenable_binlog(thd);
if (error)
{
#ifdef IN_THE_FUTURE
(void) m_file[new_part_id]->delete_last_inserted_row(new_data);
#endif
goto exit;
}
}
exit:
/*
if updating an auto_increment column, update
part_share->next_auto_inc_val if needed.
(not to be used if auto_increment on secondary field in a multi-column
index)
mysql_update does not set table->next_number_field, so we use
table->found_next_number_field instead.
Also checking that the field is marked in the write set.
*/
if (table->found_next_number_field &&
new_data == table->record[0] &&
!table->s->next_number_keypart &&
bitmap_is_set(table->write_set,
table->found_next_number_field->field_index))
{
if (!part_share->auto_inc_initialized)
info(HA_STATUS_AUTO);
set_auto_increment_if_higher(table->found_next_number_field);
}
DBUG_RETURN(error);
}
/*
Remove an existing row
SYNOPSIS
delete_row
buf Deleted row in MySQL Row Format
RETURN VALUE
>0 Error Code
0 Success
DESCRIPTION
This will delete a row. buf will contain a copy of the row to be deleted.
The server will call this right after the current row has been read
(from either a previous rnd_xxx() or index_xxx() call).
If you keep a pointer to the last row or can access a primary key it will
make doing the deletion quite a bit easier.
Keep in mind that the server does no guarentee consecutive deletions.
ORDER BY clauses can be used.
Called in sql_acl.cc and sql_udf.cc to manage internal table information.
Called in sql_delete.cc, sql_insert.cc, and sql_select.cc. In sql_select
it is used for removing duplicates while in insert it is used for REPLACE
calls.
buf is either record[0] or record[1]
*/
int ha_partition::delete_row(const uchar *buf)
{
uint32 part_id;
int error;
THD *thd= ha_thd();
DBUG_ENTER("ha_partition::delete_row");
m_err_rec= NULL;
DBUG_ASSERT(bitmap_is_subset(&m_part_info->full_part_field_set,
table->read_set));
if ((error= get_part_for_delete(buf, m_rec0, m_part_info, &part_id)))
{
DBUG_RETURN(error);
}
/* Should never call delete_row on a partition which is not read */
DBUG_ASSERT(bitmap_is_set(&(m_part_info->read_partitions), part_id));
DBUG_ASSERT(bitmap_is_set(&(m_part_info->lock_partitions), part_id));
if (!bitmap_is_set(&(m_part_info->lock_partitions), part_id))
DBUG_RETURN(HA_ERR_NOT_IN_LOCK_PARTITIONS);
/*
The protocol for deleting a row is:
1) position the handler (cursor) on the row to be deleted,
either through the last read row (rnd or index) or by rnd_pos.
2) call delete_row with the full record as argument.
This means that m_last_part should already be set to actual partition
where the row was read from. And if that is not the same as the
calculated part_id we found a misplaced row, we return an error to
notify the user that something is broken in the row distribution
between partitions! Since we don't check all rows on read, we return an
error instead of forwarding the delete to the correct (m_last_part)
partition!
Notice that HA_READ_BEFORE_WRITE_REMOVAL does not require this protocol,
so this is not supported for this engine.
TODO: change the assert in InnoDB into an error instead and make this one
an assert instead and remove the get_part_for_delete()!
*/
if (part_id != m_last_part)
{
m_err_rec= buf;
DBUG_RETURN(HA_ERR_ROW_IN_WRONG_PARTITION);
}
m_last_part= part_id;
tmp_disable_binlog(thd);
error= m_file[part_id]->ha_delete_row(buf);
reenable_binlog(thd);
DBUG_RETURN(error);
}
/*
Delete all rows in a table
SYNOPSIS
delete_all_rows()
RETURN VALUE
>0 Error Code
0 Success
DESCRIPTION
Used to delete all rows in a table. Both for cases of truncate and
for cases where the optimizer realizes that all rows will be
removed as a result of a SQL statement.
Called from item_sum.cc by Item_func_group_concat::clear(),
Item_sum_count::clear(), and Item_func_group_concat::clear().
Called from sql_delete.cc by mysql_delete().
Called from sql_select.cc by JOIN::reset().
Called from sql_union.cc by st_select_lex_unit::exec().
*/
int ha_partition::delete_all_rows()
{
int error;
uint i;
DBUG_ENTER("ha_partition::delete_all_rows");
for (i= bitmap_get_first_set(&m_part_info->read_partitions);
i < m_tot_parts;
i= bitmap_get_next_set(&m_part_info->read_partitions, i))
{
/* Can be pruned, like DELETE FROM t PARTITION (pX) */
if ((error= m_file[i]->ha_delete_all_rows()))
DBUG_RETURN(error);
}
DBUG_RETURN(0);
}
/**
Manually truncate the table.
@retval 0 Success.
@retval > 0 Error code.
*/
int ha_partition::truncate()
{
int error;
handler **file;
DBUG_ENTER("ha_partition::truncate");
/*
TRUNCATE also means resetting auto_increment. Hence, reset
it so that it will be initialized again at the next use.
*/
lock_auto_increment();
part_share->next_auto_inc_val= 0;
part_share->auto_inc_initialized= false;
unlock_auto_increment();
file= m_file;
do
{
if ((error= (*file)->ha_truncate()))
DBUG_RETURN(error);
} while (*(++file));
DBUG_RETURN(0);
}
/**
Truncate a set of specific partitions.
@remark Auto increment value will be truncated in that partition as well!
ALTER TABLE t TRUNCATE PARTITION ...
*/
int ha_partition::truncate_partition(Alter_info *alter_info, bool *binlog_stmt)
{
int error= 0;
List_iterator<partition_element> part_it(m_part_info->partitions);
uint num_parts= m_part_info->num_parts;
uint num_subparts= m_part_info->num_subparts;
uint i= 0;
DBUG_ENTER("ha_partition::truncate_partition");
/* Only binlog when it starts any call to the partitions handlers */
*binlog_stmt= false;
if (set_part_state(alter_info, m_part_info, PART_ADMIN))
DBUG_RETURN(HA_ERR_NO_PARTITION_FOUND);
/*
TRUNCATE also means resetting auto_increment. Hence, reset
it so that it will be initialized again at the next use.
*/
lock_auto_increment();
part_share->next_auto_inc_val= 0;
part_share->auto_inc_initialized= FALSE;
unlock_auto_increment();
*binlog_stmt= true;
do
{
partition_element *part_elem= part_it++;
if (part_elem->part_state == PART_ADMIN)
{
if (m_is_sub_partitioned)
{
List_iterator<partition_element>
subpart_it(part_elem->subpartitions);
partition_element *sub_elem;
uint j= 0, part;
do
{
sub_elem= subpart_it++;
part= i * num_subparts + j;
DBUG_PRINT("info", ("truncate subpartition %u (%s)",
part, sub_elem->partition_name));
if ((error= m_file[part]->ha_truncate()))
break;
sub_elem->part_state= PART_NORMAL;
} while (++j < num_subparts);
}
else
{
DBUG_PRINT("info", ("truncate partition %u (%s)", i,
part_elem->partition_name));
error= m_file[i]->ha_truncate();
}
part_elem->part_state= PART_NORMAL;
}
} while (!error && (++i < num_parts));
DBUG_RETURN(error);
}
/*
Start a large batch of insert rows
SYNOPSIS
start_bulk_insert()
rows Number of rows to insert
flags Flags to control index creation
RETURN VALUE
NONE
DESCRIPTION
rows == 0 means we will probably insert many rows
*/
void ha_partition::start_bulk_insert(ha_rows rows, uint flags)
{
DBUG_ENTER("ha_partition::start_bulk_insert");
m_bulk_inserted_rows= 0;
bitmap_clear_all(&m_bulk_insert_started);
/* use the last bit for marking if bulk_insert_started was called */
bitmap_set_bit(&m_bulk_insert_started, m_tot_parts);
DBUG_VOID_RETURN;
}
/*
Check if start_bulk_insert has been called for this partition,
if not, call it and mark it called
*/
void ha_partition::start_part_bulk_insert(THD *thd, uint part_id)
{
long old_buffer_size;
if (!bitmap_is_set(&m_bulk_insert_started, part_id) &&
bitmap_is_set(&m_bulk_insert_started, m_tot_parts))
{
DBUG_ASSERT(bitmap_is_set(&(m_part_info->lock_partitions), part_id));
old_buffer_size= thd->variables.read_buff_size;
/* Update read_buffer_size for this partition */
thd->variables.read_buff_size= estimate_read_buffer_size(old_buffer_size);
m_file[part_id]->ha_start_bulk_insert(guess_bulk_insert_rows());
bitmap_set_bit(&m_bulk_insert_started, part_id);
thd->variables.read_buff_size= old_buffer_size;
}
m_bulk_inserted_rows++;
}
/*
Estimate the read buffer size for each partition.
SYNOPSIS
ha_partition::estimate_read_buffer_size()
original_size read buffer size originally set for the server
RETURN VALUE
estimated buffer size.
DESCRIPTION
If the estimated number of rows to insert is less than 10 (but not 0)
the new buffer size is same as original buffer size.
In case of first partition of when partition function is monotonic
new buffer size is same as the original buffer size.
For rest of the partition total buffer of 10*original_size is divided
equally if number of partition is more than 10 other wise each partition
will be allowed to use original buffer size.
*/
long ha_partition::estimate_read_buffer_size(long original_size)
{
/*
If number of rows to insert is less than 10, but not 0,
return original buffer size.
*/
if (estimation_rows_to_insert && (estimation_rows_to_insert < 10))
return (original_size);
/*
If first insert/partition and monotonic partition function,
allow using buffer size originally set.
*/
if (!m_bulk_inserted_rows &&
m_part_func_monotonicity_info != NON_MONOTONIC &&
m_tot_parts > 1)
return original_size;
/*
Allow total buffer used in all partition to go up to 10*read_buffer_size.
11*read_buffer_size in case of monotonic partition function.
*/
if (m_tot_parts < 10)
return original_size;
return (original_size * 10 / m_tot_parts);
}
/*
Try to predict the number of inserts into this partition.
If less than 10 rows (including 0 which means Unknown)
just give that as a guess
If monotonic partitioning function was used
guess that 50 % of the inserts goes to the first partition
For all other cases, guess on equal distribution between the partitions
*/
ha_rows ha_partition::guess_bulk_insert_rows()
{
DBUG_ENTER("guess_bulk_insert_rows");
if (estimation_rows_to_insert < 10)
DBUG_RETURN(estimation_rows_to_insert);
/* If first insert/partition and monotonic partition function, guess 50%. */
if (!m_bulk_inserted_rows &&
m_part_func_monotonicity_info != NON_MONOTONIC &&
m_tot_parts > 1)
DBUG_RETURN(estimation_rows_to_insert / 2);
/* Else guess on equal distribution (+1 is to avoid returning 0/Unknown) */
if (m_bulk_inserted_rows < estimation_rows_to_insert)
DBUG_RETURN(((estimation_rows_to_insert - m_bulk_inserted_rows)
/ m_tot_parts) + 1);
/* The estimation was wrong, must say 'Unknown' */
DBUG_RETURN(0);
}
/*
Finish a large batch of insert rows
SYNOPSIS
end_bulk_insert()
RETURN VALUE
>0 Error code
0 Success
Note: end_bulk_insert can be called without start_bulk_insert
being called, see bug#44108.
*/
int ha_partition::end_bulk_insert()
{
int error= 0;
uint i;
DBUG_ENTER("ha_partition::end_bulk_insert");
if (!bitmap_is_set(&m_bulk_insert_started, m_tot_parts))
DBUG_RETURN(error);
for (i= bitmap_get_first_set(&m_bulk_insert_started);
i < m_tot_parts;
i= bitmap_get_next_set(&m_bulk_insert_started, i))
{
int tmp;
if ((tmp= m_file[i]->ha_end_bulk_insert()))
error= tmp;
}
bitmap_clear_all(&m_bulk_insert_started);
DBUG_RETURN(error);
}
/****************************************************************************
MODULE full table scan
****************************************************************************/
/*
Initialize engine for random reads
SYNOPSIS
ha_partition::rnd_init()
scan 0 Initialize for random reads through rnd_pos()
1 Initialize for random scan through rnd_next()
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
rnd_init() is called when the server wants the storage engine to do a
table scan or when the server wants to access data through rnd_pos.
When scan is used we will scan one handler partition at a time.
When preparing for rnd_pos we will init all handler partitions.
No extra cache handling is needed when scannning is not performed.
Before initialising we will call rnd_end to ensure that we clean up from
any previous incarnation of a table scan.
Called from filesort.cc, records.cc, sql_handler.cc, sql_select.cc,
sql_table.cc, and sql_update.cc.
*/
int ha_partition::rnd_init(bool scan)
{
int error;
uint i= 0;
uint32 part_id;
DBUG_ENTER("ha_partition::rnd_init");
/*
For operations that may need to change data, we may need to extend
read_set.
*/
if (get_lock_type() == F_WRLCK)
{
/*
If write_set contains any of the fields used in partition and
subpartition expression, we need to set all bits in read_set because
the row may need to be inserted in a different [sub]partition. In
other words update_row() can be converted into write_row(), which
requires a complete record.
*/
if (bitmap_is_overlapping(&m_part_info->full_part_field_set,
table->write_set))
bitmap_set_all(table->read_set);
else
{
/*
Some handlers only read fields as specified by the bitmap for the
read set. For partitioned handlers we always require that the
fields of the partition functions are read such that we can
calculate the partition id to place updated and deleted records.
*/
bitmap_union(table->read_set, &m_part_info->full_part_field_set);
}
}
/* Now we see what the index of our first important partition is */
DBUG_PRINT("info", ("m_part_info->read_partitions: %p",
m_part_info->read_partitions.bitmap));
part_id= bitmap_get_first_set(&(m_part_info->read_partitions));
DBUG_PRINT("info", ("m_part_spec.start_part %d", part_id));
if (MY_BIT_NONE == part_id)
{
error= 0;
goto err1;
}
/*
We have a partition and we are scanning with rnd_next
so we bump our cache
*/
DBUG_PRINT("info", ("rnd_init on partition %d", part_id));
if (scan)
{
/*
rnd_end() is needed for partitioning to reset internal data if scan
is already in use
*/
rnd_end();
late_extra_cache(part_id);
if ((error= m_file[part_id]->ha_rnd_init(scan)))
goto err;
}
else
{
for (i= part_id;
i < m_tot_parts;
i= bitmap_get_next_set(&m_part_info->read_partitions, i))
{
if ((error= m_file[i]->ha_rnd_init(scan)))
goto err;
}
}
m_scan_value= scan;
m_part_spec.start_part= part_id;
m_part_spec.end_part= m_tot_parts - 1;
DBUG_PRINT("info", ("m_scan_value=%d", m_scan_value));
DBUG_RETURN(0);
err:
/* Call rnd_end for all previously inited partitions. */
for (;
part_id < i;
part_id= bitmap_get_next_set(&m_part_info->read_partitions, part_id))
{
m_file[part_id]->ha_rnd_end();
}
err1:
m_scan_value= 2;
m_part_spec.start_part= NO_CURRENT_PART_ID;
DBUG_RETURN(error);
}
/*
End of a table scan
SYNOPSIS
rnd_end()
RETURN VALUE
>0 Error code
0 Success
*/
int ha_partition::rnd_end()
{
DBUG_ENTER("ha_partition::rnd_end");
switch (m_scan_value) {
case 2: // Error
break;
case 1:
if (NO_CURRENT_PART_ID != m_part_spec.start_part) // Table scan
{
late_extra_no_cache(m_part_spec.start_part);
m_file[m_part_spec.start_part]->ha_rnd_end();
}
break;
case 0:
uint i;
for (i= bitmap_get_first_set(&m_part_info->read_partitions);
i < m_tot_parts;
i= bitmap_get_next_set(&m_part_info->read_partitions, i))
{
m_file[i]->ha_rnd_end();
}
break;
}
m_scan_value= 2;
m_part_spec.start_part= NO_CURRENT_PART_ID;
DBUG_RETURN(0);
}
/*
read next row during full table scan (scan in random row order)
SYNOPSIS
rnd_next()
buf buffer that should be filled with data
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
This is called for each row of the table scan. When you run out of records
you should return HA_ERR_END_OF_FILE.
The Field structure for the table is the key to getting data into buf
in a manner that will allow the server to understand it.
Called from filesort.cc, records.cc, sql_handler.cc, sql_select.cc,
sql_table.cc, and sql_update.cc.
*/
int ha_partition::rnd_next(uchar *buf)
{
handler *file;
int result= HA_ERR_END_OF_FILE;
uint part_id= m_part_spec.start_part;
DBUG_ENTER("ha_partition::rnd_next");
/* upper level will increment this once again at end of call */
decrement_statistics(&SSV::ha_read_rnd_next_count);
if (NO_CURRENT_PART_ID == part_id)
{
/*
The original set of partitions to scan was empty and thus we report
the result here.
*/
goto end;
}
DBUG_ASSERT(m_scan_value == 1);
file= m_file[part_id];
while (TRUE)
{
result= file->ha_rnd_next(buf);
if (!result)
{
m_last_part= part_id;
m_part_spec.start_part= part_id;
table->status= 0;
DBUG_RETURN(0);
}
/*
if we get here, then the current partition ha_rnd_next returned failure
*/
if (result == HA_ERR_RECORD_DELETED)
continue; // Probably MyISAM
if (result != HA_ERR_END_OF_FILE)
goto end_dont_reset_start_part; // Return error
/* End current partition */
late_extra_no_cache(part_id);
DBUG_PRINT("info", ("rnd_end on partition %d", part_id));
if ((result= file->ha_rnd_end()))
break;
/* Shift to next partition */
part_id= bitmap_get_next_set(&m_part_info->read_partitions, part_id);
if (part_id >= m_tot_parts)
{
result= HA_ERR_END_OF_FILE;
break;
}
m_last_part= part_id;
m_part_spec.start_part= part_id;
file= m_file[part_id];
DBUG_PRINT("info", ("rnd_init on partition %d", part_id));
if ((result= file->ha_rnd_init(1)))
break;
late_extra_cache(part_id);
}
end:
m_part_spec.start_part= NO_CURRENT_PART_ID;
end_dont_reset_start_part:
DBUG_RETURN(result);
}
/*
Save position of current row
SYNOPSIS
position()
record Current record in MySQL Row Format
RETURN VALUE
NONE
DESCRIPTION
position() is called after each call to rnd_next() if the data needs
to be ordered. You can do something like the following to store
the position:
ha_store_ptr(ref, ref_length, current_position);
The server uses ref to store data. ref_length in the above case is
the size needed to store current_position. ref is just a byte array
that the server will maintain. If you are using offsets to mark rows, then
current_position should be the offset. If it is a primary key like in
BDB, then it needs to be a primary key.
Called from filesort.cc, sql_select.cc, sql_delete.cc and sql_update.cc.
*/
void ha_partition::position(const uchar *record)
{
handler *file= m_file[m_last_part];
uint pad_length;
DBUG_ASSERT(bitmap_is_set(&(m_part_info->read_partitions), m_last_part));
DBUG_ENTER("ha_partition::position");
file->position(record);
int2store(ref, m_last_part);
memcpy((ref + PARTITION_BYTES_IN_POS), file->ref, file->ref_length);
pad_length= m_ref_length - PARTITION_BYTES_IN_POS - file->ref_length;
if (pad_length)
memset((ref + PARTITION_BYTES_IN_POS + file->ref_length), 0, pad_length);
DBUG_VOID_RETURN;
}
/*
Read row using position
SYNOPSIS
rnd_pos()
out:buf Row read in MySQL Row Format
position Position of read row
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
This is like rnd_next, but you are given a position to use
to determine the row. The position will be of the type that you stored in
ref. You can use ha_get_ptr(pos,ref_length) to retrieve whatever key
or position you saved when position() was called.
Called from filesort.cc records.cc sql_insert.cc sql_select.cc
sql_update.cc.
*/
int ha_partition::rnd_pos(uchar * buf, uchar *pos)
{
uint part_id;
handler *file;
DBUG_ENTER("ha_partition::rnd_pos");
decrement_statistics(&SSV::ha_read_rnd_count);
part_id= uint2korr((const uchar *) pos);
DBUG_ASSERT(part_id < m_tot_parts);
file= m_file[part_id];
DBUG_ASSERT(bitmap_is_set(&(m_part_info->read_partitions), part_id));
m_last_part= part_id;
DBUG_RETURN(file->ha_rnd_pos(buf, (pos + PARTITION_BYTES_IN_POS)));
}
/*
Read row using position using given record to find
SYNOPSIS
rnd_pos_by_record()
record Current record in MySQL Row Format
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
this works as position()+rnd_pos() functions, but does some extra work,
calculating m_last_part - the partition to where the 'record'
should go.
called from replication (log_event.cc)
*/
int ha_partition::rnd_pos_by_record(uchar *record)
{
DBUG_ENTER("ha_partition::rnd_pos_by_record");
if (unlikely(get_part_for_delete(record, m_rec0, m_part_info, &m_last_part)))
DBUG_RETURN(1);
DBUG_RETURN(handler::rnd_pos_by_record(record));
}
/****************************************************************************
MODULE index scan
****************************************************************************/
/*
Positions an index cursor to the index specified in the handle. Fetches the
row if available. If the key value is null, begin at the first key of the
index.
There are loads of optimisations possible here for the partition handler.
The same optimisations can also be checked for full table scan although
only through conditions and not from index ranges.
Phase one optimisations:
Check if the fields of the partition function are bound. If so only use
the single partition it becomes bound to.
Phase two optimisations:
If it can be deducted through range or list partitioning that only a
subset of the partitions are used, then only use those partitions.
*/
/**
Setup the ordered record buffer and the priority queue.
*/
bool ha_partition::init_record_priority_queue()
{
DBUG_ENTER("ha_partition::init_record_priority_queue");
DBUG_ASSERT(!m_ordered_rec_buffer);
/*
Initialize the ordered record buffer.
*/
if (!m_ordered_rec_buffer)
{
uint alloc_len;
uint used_parts= bitmap_bits_set(&m_part_info->read_partitions);
/* Allocate record buffer for each used partition. */
m_priority_queue_rec_len= m_rec_length + PARTITION_BYTES_IN_POS;
if (!m_using_extended_keys)
m_priority_queue_rec_len += m_file[0]->ref_length;
alloc_len= used_parts * m_priority_queue_rec_len;
/* Allocate a key for temporary use when setting up the scan. */
alloc_len+= table_share->max_key_length;
if (!(m_ordered_rec_buffer= (uchar*)my_malloc(alloc_len, MYF(MY_WME))))
DBUG_RETURN(true);
/*
We set-up one record per partition and each record has 2 bytes in
front where the partition id is written. This is used by ordered
index_read.
We also set-up a reference to the first record for temporary use in
setting up the scan.
*/
char *ptr= (char*) m_ordered_rec_buffer;
uint i;
for (i= bitmap_get_first_set(&m_part_info->read_partitions);
i < m_tot_parts;
i= bitmap_get_next_set(&m_part_info->read_partitions, i))
{
DBUG_PRINT("info", ("init rec-buf for part %u", i));
int2store(ptr, i);
ptr+= m_priority_queue_rec_len;
}
m_start_key.key= (const uchar*)ptr;
/* Initialize priority queue, initialized to reading forward. */
int (*cmp_func)(void *, uchar *, uchar *);
void *cmp_arg;
if (!m_using_extended_keys)
{
cmp_func= cmp_key_rowid_part_id;
cmp_arg= (void*)this;
}
else
{
cmp_func= cmp_key_part_id;
cmp_arg= (void*)m_curr_key_info;
}
if (init_queue(&m_queue, used_parts, 0, 0, cmp_func, cmp_arg, 0, 0))
{
my_free(m_ordered_rec_buffer);
m_ordered_rec_buffer= NULL;
DBUG_RETURN(true);
}
}
DBUG_RETURN(false);
}
/**
Destroy the ordered record buffer and the priority queue.
*/
void ha_partition::destroy_record_priority_queue()
{
DBUG_ENTER("ha_partition::destroy_record_priority_queue");
if (m_ordered_rec_buffer)
{
delete_queue(&m_queue);
my_free(m_ordered_rec_buffer);
m_ordered_rec_buffer= NULL;
}
DBUG_VOID_RETURN;
}
/*
Initialize handler before start of index scan
SYNOPSIS
index_init()
inx Index number
sorted Is rows to be returned in sorted order
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
index_init is always called before starting index scans (except when
starting through index_read_idx and using read_range variants).
*/
int ha_partition::index_init(uint inx, bool sorted)
{
int error= 0;
uint i;
DBUG_ENTER("ha_partition::index_init");
DBUG_PRINT("info", ("inx %u sorted %u", inx, sorted));
active_index= inx;
m_part_spec.start_part= NO_CURRENT_PART_ID;
m_start_key.length= 0;
m_ordered= sorted;
m_ordered_scan_ongoing= FALSE;
m_curr_key_info[0]= table->key_info+inx;
if (m_pkey_is_clustered && table->s->primary_key != MAX_KEY)
{
/*
if PK is clustered, then the key cmp must use the pk to
differentiate between equal key in given index.
*/
DBUG_PRINT("info", ("Clustered pk, using pk as secondary cmp"));
m_curr_key_info[1]= table->key_info+table->s->primary_key;
m_curr_key_info[2]= NULL;
m_using_extended_keys= TRUE;
}
else
{
m_curr_key_info[1]= NULL;
m_using_extended_keys= FALSE;
}
if (init_record_priority_queue())
DBUG_RETURN(HA_ERR_OUT_OF_MEM);
/*
Some handlers only read fields as specified by the bitmap for the
read set. For partitioned handlers we always require that the
fields of the partition functions are read such that we can
calculate the partition id to place updated and deleted records.
But this is required for operations that may need to change data only.
*/
if (get_lock_type() == F_WRLCK)
bitmap_union(table->read_set, &m_part_info->full_part_field_set);
if (sorted)
{
/*
An ordered scan is requested. We must make sure all fields of the
used index are in the read set, as partitioning requires them for
sorting (see ha_partition::handle_ordered_index_scan).
The SQL layer may request an ordered index scan without having index
fields in the read set when
- it needs to do an ordered scan over an index prefix.
- it evaluates ORDER BY with SELECT COUNT(*) FROM t1.
TODO: handle COUNT(*) queries via unordered scan.
*/
KEY **key_info= m_curr_key_info;
do
{
for (i= 0; i < (*key_info)->user_defined_key_parts; i++)
bitmap_set_bit(table->read_set,
(*key_info)->key_part[i].field->field_index);
} while (*(++key_info));
}
for (i= bitmap_get_first_set(&m_part_info->read_partitions);
i < m_tot_parts;
i= bitmap_get_next_set(&m_part_info->read_partitions, i))
{
if ((error= m_file[i]->ha_index_init(inx, sorted)))
goto err;
DBUG_EXECUTE_IF("ha_partition_fail_index_init", {
i++;
error= HA_ERR_NO_PARTITION_FOUND;
goto err;
});
}
err:
if (error)
{
/* End the previously initialized indexes. */
uint j;
for (j= bitmap_get_first_set(&m_part_info->read_partitions);
j < i;
j= bitmap_get_next_set(&m_part_info->read_partitions, j))
{
(void) m_file[j]->ha_index_end();
}
destroy_record_priority_queue();
}
DBUG_RETURN(error);
}
/*
End of index scan
SYNOPSIS
index_end()
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
index_end is called at the end of an index scan to clean up any
things needed to clean up.
*/
int ha_partition::index_end()
{
int error= 0;
uint i;
DBUG_ENTER("ha_partition::index_end");
active_index= MAX_KEY;
m_part_spec.start_part= NO_CURRENT_PART_ID;
for (i= bitmap_get_first_set(&m_part_info->read_partitions);
i < m_tot_parts;
i= bitmap_get_next_set(&m_part_info->read_partitions, i))
{
int tmp;
if ((tmp= m_file[i]->ha_index_end()))
error= tmp;
}
destroy_record_priority_queue();
DBUG_RETURN(error);
}
/*
Read one record in an index scan and start an index scan
SYNOPSIS
index_read_map()
buf Read row in MySQL Row Format
key Key parts in consecutive order
keypart_map Which part of key is used
find_flag What type of key condition is used
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
index_read_map starts a new index scan using a start key. The MySQL Server
will check the end key on its own. Thus to function properly the
partitioned handler need to ensure that it delivers records in the sort
order of the MySQL Server.
index_read_map can be restarted without calling index_end on the previous
index scan and without calling index_init. In this case the index_read_map
is on the same index as the previous index_scan. This is particularly
used in conjuntion with multi read ranges.
*/
int ha_partition::index_read_map(uchar *buf, const uchar *key,
key_part_map keypart_map,
enum ha_rkey_function find_flag)
{
DBUG_ENTER("ha_partition::index_read_map");
decrement_statistics(&SSV::ha_read_key_count);
end_range= 0;
m_index_scan_type= partition_index_read;
m_start_key.key= key;
m_start_key.keypart_map= keypart_map;
m_start_key.flag= find_flag;
DBUG_RETURN(common_index_read(buf, TRUE));
}
/* Compare two part_no partition numbers */
static int cmp_part_ids(uchar *ref1, uchar *ref2)
{
/* The following was taken from ha_partition::cmp_ref */
my_ptrdiff_t diff1= ref2[1] - ref1[1];
my_ptrdiff_t diff2= ref2[0] - ref1[0];
if (!diff1 && !diff2)
return 0;
if (diff1 > 0)
return(-1);
if (diff1 < 0)
return(+1);
if (diff2 > 0)
return(-1);
return(+1);
}
/*
@brief
Provide ordering by (key_value, part_no).
*/
extern "C" int cmp_key_part_id(void *key_p, uchar *ref1, uchar *ref2)
{
int res;
if ((res= key_rec_cmp(key_p, ref1 + PARTITION_BYTES_IN_POS,
ref2 + PARTITION_BYTES_IN_POS)))
{
return res;
}
return cmp_part_ids(ref1, ref2);
}
/*
@brief
Provide ordering by (key_value, underying_table_rowid, part_no).
*/
extern "C" int cmp_key_rowid_part_id(void *ptr, uchar *ref1, uchar *ref2)
{
ha_partition *file= (ha_partition*)ptr;
int res;
if ((res= key_rec_cmp(file->m_curr_key_info, ref1 + PARTITION_BYTES_IN_POS,
ref2 + PARTITION_BYTES_IN_POS)))
{
return res;
}
if ((res= file->m_file[0]->cmp_ref(ref1 + PARTITION_BYTES_IN_POS + file->m_rec_length,
ref2 + PARTITION_BYTES_IN_POS + file->m_rec_length)))
{
return res;
}
return cmp_part_ids(ref1, ref2);
}
/**
Common routine for a number of index_read variants
@param buf Buffer where the record should be returned.
@param have_start_key TRUE <=> the left endpoint is available, i.e.
we're in index_read call or in read_range_first
call and the range has left endpoint.
FALSE <=> there is no left endpoint (we're in
read_range_first() call and the range has no left
endpoint).
@return Operation status
@retval 0 OK
@retval HA_ERR_END_OF_FILE Whole index scanned, without finding the record.
@retval HA_ERR_KEY_NOT_FOUND Record not found, but index cursor positioned.
@retval other error code.
@details
Start scanning the range (when invoked from read_range_first()) or doing
an index lookup (when invoked from index_read_XXX):
- If possible, perform partition selection
- Find the set of partitions we're going to use
- Depending on whether we need ordering:
NO: Get the first record from first used partition (see
handle_unordered_scan_next_partition)
YES: Fill the priority queue and get the record that is the first in
the ordering
*/
int ha_partition::common_index_read(uchar *buf, bool have_start_key)
{
int error;
uint UNINIT_VAR(key_len); /* used if have_start_key==TRUE */
bool reverse_order= FALSE;
DBUG_ENTER("ha_partition::common_index_read");
DBUG_PRINT("info", ("m_ordered %u m_ordered_scan_ong %u",
m_ordered, m_ordered_scan_ongoing));
if (have_start_key)
{
m_start_key.length= key_len= calculate_key_len(table, active_index,
m_start_key.key,
m_start_key.keypart_map);
DBUG_PRINT("info", ("have_start_key map %lu find_flag %u len %u",
m_start_key.keypart_map, m_start_key.flag, key_len));
DBUG_ASSERT(key_len);
}
if ((error= partition_scan_set_up(buf, have_start_key)))
{
DBUG_RETURN(error);
}
if (have_start_key &&
(m_start_key.flag == HA_READ_PREFIX_LAST ||
m_start_key.flag == HA_READ_PREFIX_LAST_OR_PREV ||
m_start_key.flag == HA_READ_BEFORE_KEY))
{
reverse_order= TRUE;
m_ordered_scan_ongoing= TRUE;
}
DBUG_PRINT("info", ("m_ordered %u m_o_scan_ong %u have_start_key %u",
m_ordered, m_ordered_scan_ongoing, have_start_key));
if (!m_ordered_scan_ongoing)
{
/*
We use unordered index scan when read_range is used and flag
is set to not use ordered.
We also use an unordered index scan when the number of partitions to
scan is only one.
The unordered index scan will use the partition set created.
*/
DBUG_PRINT("info", ("doing unordered scan"));
error= handle_unordered_scan_next_partition(buf);
}
else
{
/*
In all other cases we will use the ordered index scan. This will use
the partition set created by the get_partition_set method.
*/
error= handle_ordered_index_scan(buf, reverse_order);
}
DBUG_RETURN(error);
}
/*
Start an index scan from leftmost record and return first record
SYNOPSIS
index_first()
buf Read row in MySQL Row Format
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
index_first() asks for the first key in the index.
This is similar to index_read except that there is no start key since
the scan starts from the leftmost entry and proceeds forward with
index_next.
Called from opt_range.cc, opt_sum.cc, sql_handler.cc,
and sql_select.cc.
*/
int ha_partition::index_first(uchar * buf)
{
DBUG_ENTER("ha_partition::index_first");
decrement_statistics(&SSV::ha_read_first_count);
end_range= 0;
m_index_scan_type= partition_index_first;
DBUG_RETURN(common_first_last(buf));
}
/*
Start an index scan from rightmost record and return first record
SYNOPSIS
index_last()
buf Read row in MySQL Row Format
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
index_last() asks for the last key in the index.
This is similar to index_read except that there is no start key since
the scan starts from the rightmost entry and proceeds forward with
index_prev.
Called from opt_range.cc, opt_sum.cc, sql_handler.cc,
and sql_select.cc.
*/
int ha_partition::index_last(uchar * buf)
{
DBUG_ENTER("ha_partition::index_last");
decrement_statistics(&SSV::ha_read_last_count);
m_index_scan_type= partition_index_last;
DBUG_RETURN(common_first_last(buf));
}
/*
Common routine for index_first/index_last
SYNOPSIS
ha_partition::common_first_last()
see index_first for rest
*/
int ha_partition::common_first_last(uchar *buf)
{
int error;
if ((error= partition_scan_set_up(buf, FALSE)))
return error;
if (!m_ordered_scan_ongoing &&
m_index_scan_type != partition_index_last)
return handle_unordered_scan_next_partition(buf);
return handle_ordered_index_scan(buf, FALSE);
}
/*
Optimization of the default implementation to take advantage of dynamic
partition pruning.
*/
int ha_partition::index_read_idx_map(uchar *buf, uint index,
const uchar *key,
key_part_map keypart_map,
enum ha_rkey_function find_flag)
{
int error= HA_ERR_KEY_NOT_FOUND;
DBUG_ENTER("ha_partition::index_read_idx_map");
if (find_flag == HA_READ_KEY_EXACT)
{
uint part;
m_start_key.key= key;
m_start_key.keypart_map= keypart_map;
m_start_key.flag= find_flag;
m_start_key.length= calculate_key_len(table, index, m_start_key.key,
m_start_key.keypart_map);
get_partition_set(table, buf, index, &m_start_key, &m_part_spec);
/*
We have either found exactly 1 partition
(in which case start_part == end_part)
or no matching partitions (start_part > end_part)
*/
DBUG_ASSERT(m_part_spec.start_part >= m_part_spec.end_part);
/* The start part is must be marked as used. */
DBUG_ASSERT(m_part_spec.start_part > m_part_spec.end_part ||
bitmap_is_set(&(m_part_info->read_partitions),
m_part_spec.start_part));
for (part= m_part_spec.start_part;
part <= m_part_spec.end_part;
part= bitmap_get_next_set(&m_part_info->read_partitions, part))
{
error= m_file[part]->ha_index_read_idx_map(buf, index, key,
keypart_map, find_flag);
if (error != HA_ERR_KEY_NOT_FOUND &&
error != HA_ERR_END_OF_FILE)
break;
}
if (part <= m_part_spec.end_part)
m_last_part= part;
}
else
{
/*
If not only used with READ_EXACT, we should investigate if possible
to optimize for other find_flag's as well.
*/
DBUG_ASSERT(0);
/* fall back on the default implementation */
error= handler::index_read_idx_map(buf, index, key, keypart_map, find_flag);
}
DBUG_RETURN(error);
}
/*
Read next record in a forward index scan
SYNOPSIS
index_next()
buf Read row in MySQL Row Format
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
Used to read forward through the index.
*/
int ha_partition::index_next(uchar * buf)
{
DBUG_ENTER("ha_partition::index_next");
decrement_statistics(&SSV::ha_read_next_count);
/*
TODO(low priority):
If we want partition to work with the HANDLER commands, we
must be able to do index_last() -> index_prev() -> index_next()
and if direction changes, we must step back those partitions in
the record queue so we don't return a value from the wrong direction.
*/
DBUG_ASSERT(m_index_scan_type != partition_index_last);
if (!m_ordered_scan_ongoing)
{
DBUG_RETURN(handle_unordered_next(buf, FALSE));
}
DBUG_RETURN(handle_ordered_next(buf, FALSE));
}
/*
Read next record special
SYNOPSIS
index_next_same()
buf Read row in MySQL Row Format
key Key
keylen Length of key
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
This routine is used to read the next but only if the key is the same
as supplied in the call.
*/
int ha_partition::index_next_same(uchar *buf, const uchar *key, uint keylen)
{
DBUG_ENTER("ha_partition::index_next_same");
decrement_statistics(&SSV::ha_read_next_count);
DBUG_ASSERT(keylen == m_start_key.length);
DBUG_ASSERT(m_index_scan_type != partition_index_last);
if (!m_ordered_scan_ongoing)
DBUG_RETURN(handle_unordered_next(buf, TRUE));
DBUG_RETURN(handle_ordered_next(buf, TRUE));
}
int ha_partition::index_read_last_map(uchar *buf,
const uchar *key,
key_part_map keypart_map)
{
DBUG_ENTER("ha_partition::index_read_last_map");
m_ordered= true; // Safety measure
end_range= NULL;
m_index_scan_type= partition_index_read_last;
m_start_key.key= key;
m_start_key.keypart_map= keypart_map;
m_start_key.flag= HA_READ_PREFIX_LAST;
DBUG_RETURN(common_index_read(buf, true));
}
/*
Read next record when performing index scan backwards
SYNOPSIS
index_prev()
buf Read row in MySQL Row Format
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
Used to read backwards through the index.
*/
int ha_partition::index_prev(uchar * buf)
{
DBUG_ENTER("ha_partition::index_prev");
decrement_statistics(&SSV::ha_read_prev_count);
/* TODO: read comment in index_next */
DBUG_ASSERT(m_index_scan_type != partition_index_first);
DBUG_RETURN(handle_ordered_prev(buf));
}
/*
Start a read of one range with start and end key
SYNOPSIS
read_range_first()
start_key Specification of start key
end_key Specification of end key
eq_range_arg Is it equal range
sorted Should records be returned in sorted order
RETURN VALUE
>0 Error code
0 Success
DESCRIPTION
We reimplement read_range_first since we don't want the compare_key
check at the end. This is already performed in the partition handler.
read_range_next is very much different due to that we need to scan
all underlying handlers.
*/
int ha_partition::read_range_first(const key_range *start_key,
const key_range *end_key,
bool eq_range_arg, bool sorted)
{
int error;
DBUG_ENTER("ha_partition::read_range_first");
m_ordered= sorted;
eq_range= eq_range_arg;
set_end_range(end_key);
range_key_part= m_curr_key_info[0]->key_part;
if (start_key)
m_start_key= *start_key;
else
m_start_key.key= NULL;
m_index_scan_type= partition_read_range;
error= common_index_read(m_rec0, MY_TEST(start_key));
DBUG_RETURN(error);
}
/*
Read next record in read of a range with start and end key
SYNOPSIS
read_range_next()
RETURN VALUE
>0 Error code
0 Success
*/
int ha_partition::read_range_next()
{
DBUG_ENTER("ha_partition::read_range_next");
if (m_ordered_scan_ongoing)
{
DBUG_RETURN(handle_ordered_next(table->record[0], eq_range));
}
DBUG_RETURN(handle_unordered_next(table->record[0], eq_range));
}
/*
Common routine to set up index scans
SYNOPSIS
ha_partition::partition_scan_set_up()
buf Buffer to later return record in (this function
needs it to calculcate partitioning function
values)
idx_read_flag TRUE <=> m_start_key has range start endpoint which
probably can be used to determine the set of partitions
to scan.
FALSE <=> there is no start endpoint.
DESCRIPTION
Find out which partitions we'll need to read when scanning the specified
range.
If we need to scan only one partition, set m_ordered_scan_ongoing=FALSE
as we will not need to do merge ordering.
RETURN VALUE
>0 Error code
0 Success
*/
int ha_partition::partition_scan_set_up(uchar * buf, bool idx_read_flag)
{
DBUG_ENTER("ha_partition::partition_scan_set_up");
if (idx_read_flag)
get_partition_set(table,buf,active_index,&m_start_key,&m_part_spec);
else
{
m_part_spec.start_part= 0;
m_part_spec.end_part= m_tot_parts - 1;
}
if (m_part_spec.start_part > m_part_spec.end_part)
{
/*
We discovered a partition set but the set was empty so we report
key not found.
*/
DBUG_PRINT("info", ("scan with no partition to scan"));
DBUG_RETURN(HA_ERR_END_OF_FILE);
}
if (m_part_spec.start_part == m_part_spec.end_part)
{
/*
We discovered a single partition to scan, this never needs to be
performed using the ordered index scan.
*/
DBUG_PRINT("info", ("index scan using the single partition %d",
m_part_spec.start_part));
m_ordered_scan_ongoing= FALSE;
}
else
{
/*
Set m_ordered_scan_ongoing according how the scan should be done
Only exact partitions are discovered atm by get_partition_set.
Verify this, also bitmap must have at least one bit set otherwise
the result from this table is the empty set.
*/
uint start_part= bitmap_get_first_set(&(m_part_info->read_partitions));
if (start_part == MY_BIT_NONE)
{
DBUG_PRINT("info", ("scan with no partition to scan"));
DBUG_RETURN(HA_ERR_END_OF_FILE);
}
if (start_part > m_part_spec.start_part)
m_part_spec.start_part= start_part;
DBUG_ASSERT(m_part_spec.start_part < m_tot_parts);
m_ordered_scan_ongoing= m_ordered;
}
DBUG_ASSERT(m_part_spec.start_part < m_tot_parts &&
m_part_spec.end_part < m_tot_parts);
DBUG_RETURN(0);
}
/****************************************************************************
Unordered Index Scan Routines
****************************************************************************/
/*
Common routine to handle index_next with unordered results
SYNOPSIS
handle_unordered_next()
out:buf Read row in MySQL Row Format
next_same Called from index_next_same
RETURN VALUE
HA_ERR_END_OF_FILE End of scan
0 Success
other Error code
DESCRIPTION
These routines are used to scan partitions without considering order.
This is performed in two situations.
1) In read_multi_range this is the normal case
2) When performing any type of index_read, index_first, index_last where
all fields in the partition function is bound. In this case the index
scan is performed on only one partition and thus it isn't necessary to
perform any sort.
*/
int ha_partition::handle_unordered_next(uchar *buf, bool is_next_same)
{
handler *file;
int error;
DBUG_ENTER("ha_partition::handle_unordered_next");
if (m_part_spec.start_part >= m_tot_parts)
{
/* Should never happen! */
DBUG_ASSERT(0);
DBUG_RETURN(HA_ERR_END_OF_FILE);
}
file= m_file[m_part_spec.start_part];
/*
We should consider if this should be split into three functions as
partition_read_range is_next_same are always local constants
*/
if (m_index_scan_type == partition_read_range)
{
if (!(error= file->read_range_next()))
{
m_last_part= m_part_spec.start_part;
DBUG_RETURN(0);
}
}
else if (is_next_same)
{
if (!(error= file->ha_index_next_same(buf, m_start_key.key,
m_start_key.length)))
{
m_last_part= m_part_spec.start_part;
DBUG_RETURN(0);
}
}
else
{
if (!(error= file->ha_index_next(buf)))
{
m_last_part= m_part_spec.start_part;
DBUG_RETURN(0); // Row was in range
}
}
if (error == HA_ERR_END_OF_FILE)
{
m_part_spec.start_part++; // Start using next part
error= handle_unordered_scan_next_partition(buf);
}
DBUG_RETURN(error);
}
/*
Handle index_next when changing to new partition
SYNOPSIS
handle_unordered_scan_next_partition()
buf Read row in MySQL Row Format
RETURN VALUE
HA_ERR_END_OF_FILE End of scan
0 Success
other Error code
DESCRIPTION
This routine is used to start the index scan on the next partition.
Both initial start and after completing scan on one partition.
*/
int ha_partition::handle_unordered_scan_next_partition(uchar * buf)
{
uint i= m_part_spec.start_part;
int saved_error= HA_ERR_END_OF_FILE;
DBUG_ENTER("ha_partition::handle_unordered_scan_next_partition");
if (i)
i= bitmap_get_next_set(&m_part_info->read_partitions, i - 1);
else
i= bitmap_get_first_set(&m_part_info->read_partitions);
for (;
i <= m_part_spec.end_part;
i= bitmap_get_next_set(&m_part_info->read_partitions, i))
{
int error;
handler *file= m_file[i];
m_part_spec.start_part= i;
switch (m_index_scan_type) {
case partition_read_range:
DBUG_PRINT("info", ("read_range_first on partition %d", i));
error= file->read_range_first(m_start_key.key? &m_start_key: NULL,
end_range, eq_range, FALSE);
break;
case partition_index_read:
DBUG_PRINT("info", ("index_read on partition %d", i));
error= file->ha_index_read_map(buf, m_start_key.key,
m_start_key.keypart_map,
m_start_key.flag);
break;
case partition_index_first:
DBUG_PRINT("info", ("index_first on partition %d", i));
error= file->ha_index_first(buf);
break;
case partition_index_first_unordered:
/*
We perform a scan without sorting and this means that we
should not use the index_first since not all handlers
support it and it is also unnecessary to restrict sort
order.
*/
DBUG_PRINT("info", ("read_range_first on partition %d", i));
table->record[0]= buf;
error= file->read_range_first(0, end_range, eq_range, 0);
table->record[0]= m_rec0;
break;
default:
DBUG_ASSERT(FALSE);
DBUG_RETURN(1);
}
if (!error)
{
m_last_part= i;
DBUG_RETURN(0);
}
if ((error != HA_ERR_END_OF_FILE) && (error != HA_ERR_KEY_NOT_FOUND))
DBUG_RETURN(error);
/*
If HA_ERR_KEY_NOT_FOUND, we must return that error instead of
HA_ERR_END_OF_FILE, to be able to continue search.
*/
if (saved_error != HA_ERR_KEY_NOT_FOUND)
saved_error= error;
DBUG_PRINT("info", ("END_OF_FILE/KEY_NOT_FOUND on partition %d", i));
}
if (saved_error == HA_ERR_END_OF_FILE)
m_part_spec.start_part= NO_CURRENT_PART_ID;
DBUG_RETURN(saved_error);
}
/**
Common routine to start index scan with ordered results.
@param[out] buf Read row in MySQL Row Format
@return Operation status
@retval HA_ERR_END_OF_FILE End of scan
@retval HA_ERR_KEY_NOT_FOUNE End of scan
@retval 0 Success
@retval other Error code
@details
This part contains the logic to handle index scans that require ordered
output. This includes all except those started by read_range_first with
the flag ordered set to FALSE. Thus most direct index_read and all
index_first and index_last.
We implement ordering by keeping one record plus a key buffer for each
partition. Every time a new entry is requested we will fetch a new
entry from the partition that is currently not filled with an entry.
Then the entry is put into its proper sort position.
Returning a record is done by getting the top record, copying the
record to the request buffer and setting the partition as empty on
entries.
*/
int ha_partition::handle_ordered_index_scan(uchar *buf, bool reverse_order)
{
uint i;
uint j= queue_first_element(&m_queue);
bool found= FALSE;
uchar *part_rec_buf_ptr= m_ordered_rec_buffer;
int saved_error= HA_ERR_END_OF_FILE;
DBUG_ENTER("ha_partition::handle_ordered_index_scan");
if (m_key_not_found)
{
m_key_not_found= false;
bitmap_clear_all(&m_key_not_found_partitions);
}
m_top_entry= NO_CURRENT_PART_ID;
queue_remove_all(&m_queue);
DBUG_ASSERT(bitmap_is_set(&m_part_info->read_partitions,
m_part_spec.start_part));
/*
Position part_rec_buf_ptr to point to the first used partition >=
start_part. There may be partitions marked by used_partitions,
but is before start_part. These partitions has allocated record buffers
but is dynamically pruned, so those buffers must be skipped.
*/
for (i= bitmap_get_first_set(&m_part_info->read_partitions);
i < m_part_spec.start_part;
i= bitmap_get_next_set(&m_part_info->read_partitions, i))
{
part_rec_buf_ptr+= m_priority_queue_rec_len;
}
DBUG_PRINT("info", ("m_part_spec.start_part %u first_used_part %u",
m_part_spec.start_part, i));
for (/* continue from above */ ;
i <= m_part_spec.end_part;
i= bitmap_get_next_set(&m_part_info->read_partitions, i))
{
DBUG_PRINT("info", ("reading from part %u (scan_type: %u)",
i, m_index_scan_type));
DBUG_ASSERT(i == uint2korr(part_rec_buf_ptr));
uchar *rec_buf_ptr= part_rec_buf_ptr + PARTITION_BYTES_IN_POS;
int error;
handler *file= m_file[i];
switch (m_index_scan_type) {
case partition_index_read:
error= file->ha_index_read_map(rec_buf_ptr,
m_start_key.key,
m_start_key.keypart_map,
m_start_key.flag);
break;
case partition_index_first:
error= file->ha_index_first(rec_buf_ptr);
reverse_order= FALSE;
break;
case partition_index_last:
error= file->ha_index_last(rec_buf_ptr);
reverse_order= TRUE;
break;
case partition_read_range:
{
/*
This can only read record to table->record[0], as it was set when
the table was being opened. We have to memcpy data ourselves.
*/
error= file->read_range_first(m_start_key.key? &m_start_key: NULL,
end_range, eq_range, TRUE);
memcpy(rec_buf_ptr, table->record[0], m_rec_length);
reverse_order= FALSE;
break;
}
default:
DBUG_ASSERT(FALSE);
DBUG_RETURN(HA_ERR_END_OF_FILE);
}
if (!error)
{
found= TRUE;
if (!m_using_extended_keys)
{
file->position(rec_buf_ptr);
memcpy(rec_buf_ptr + m_rec_length, file->ref, file->ref_length);
}
/*
Initialize queue without order first, simply insert
*/
queue_element(&m_queue, j++)= part_rec_buf_ptr;
}
else if (error != HA_ERR_KEY_NOT_FOUND && error != HA_ERR_END_OF_FILE)
{
DBUG_RETURN(error);
}
else if (error == HA_ERR_KEY_NOT_FOUND)
{
DBUG_PRINT("info", ("HA_ERR_KEY_NOT_FOUND from partition %u", i));
bitmap_set_bit(&m_key_not_found_partitions, i);
m_key_not_found= true;
saved_error= error;
}
part_rec_buf_ptr+= m_priority_queue_rec_len;
}
if (found)
{
/*
We found at least one partition with data, now sort all entries and
after that read the first entry and copy it to the buffer to return in.
*/
queue_set_max_at_top(&m_queue, reverse_order);
queue_set_cmp_arg(&m_queue, m_using_extended_keys? m_curr_key_info : (void*)this);
m_queue.elements= j - queue_first_element(&m_queue);
queue_fix(&m_queue);
return_top_record(buf);
table->status= 0;
DBUG_PRINT("info", ("Record returned from partition %d", m_top_entry));
DBUG_RETURN(0);
}
DBUG_RETURN(saved_error);
}
/*
Return the top record in sort order
SYNOPSIS
return_top_record()
out:buf Row returned in MySQL Row Format
RETURN VALUE
NONE
*/
void ha_partition::return_top_record(uchar *buf)
{
uint part_id;
uchar *key_buffer= queue_top(&m_queue);
uchar *rec_buffer= key_buffer + PARTITION_BYTES_IN_POS;
part_id= uint2korr(key_buffer);
memcpy(buf, rec_buffer, m_rec_length);
m_last_part= part_id;
m_top_entry= part_id;
}
/**
Add index_next/prev from partitions without exact match.
If there where any partitions that returned HA_ERR_KEY_NOT_FOUND when
ha_index_read_map was done, those partitions must be included in the
following index_next/prev call.
*/
int ha_partition::handle_ordered_index_scan_key_not_found()
{
int error;
uint i, old_elements= m_queue.elements;
uchar *part_buf= m_ordered_rec_buffer;
uchar *curr_rec_buf= NULL;
DBUG_ENTER("ha_partition::handle_ordered_index_scan_key_not_found");
DBUG_ASSERT(m_key_not_found);
/*
Loop over all used partitions to get the correct offset
into m_ordered_rec_buffer.
*/
for (i= bitmap_get_first_set(&m_part_info->read_partitions);
i < m_tot_parts;
i= bitmap_get_next_set(&m_part_info->read_partitions, i))
{
if (bitmap_is_set(&m_key_not_found_partitions, i))
{
/*
This partition is used and did return HA_ERR_KEY_NOT_FOUND
in index_read_map.
*/
curr_rec_buf= part_buf + PARTITION_BYTES_IN_POS;
error= m_file[i]->ha_index_next(curr_rec_buf);
/* HA_ERR_KEY_NOT_FOUND is not allowed from index_next! */
DBUG_ASSERT(error != HA_ERR_KEY_NOT_FOUND);
if (!error)
queue_insert(&m_queue, part_buf);
else if (error != HA_ERR_END_OF_FILE && error != HA_ERR_KEY_NOT_FOUND)
DBUG_RETURN(error);
}
part_buf += m_priority_queue_rec_len;
}
DBUG_ASSERT(curr_rec_buf);
bitmap_clear_all(&m_key_not_found_partitions);
m_key_not_found= false;
if (m_queue.elements > old_elements)
{
/* Update m_top_entry, which may have changed. */
uchar *key_buffer= queue_top(&m_queue);
m_top_entry= uint2korr(key_buffer);
}
DBUG_RETURN(0);
}
/*
Common routine to handle index_next with ordered results
SYNOPSIS
handle_ordered_next()
out:buf Read row in MySQL Row Format
next_same Called from index_next_same
RETURN VALUE
HA_ERR_END_OF_FILE End of scan
0 Success
other Error code
*/
int ha_partition::handle_ordered_next(uchar *buf, bool is_next_same)
{
int error;
uint part_id= m_top_entry;
uchar *rec_buf= queue_top(&m_queue) + PARTITION_BYTES_IN_POS;
handler *file;
DBUG_ENTER("ha_partition::handle_ordered_next");
if (m_key_not_found)
{
if (is_next_same)
{
/* Only rows which match the key. */
m_key_not_found= false;
bitmap_clear_all(&m_key_not_found_partitions);
}
else
{
/* There are partitions not included in the index record queue. */
uint old_elements= m_queue.elements;
if ((error= handle_ordered_index_scan_key_not_found()))
DBUG_RETURN(error);
/*
If the queue top changed, i.e. one of the partitions that gave
HA_ERR_KEY_NOT_FOUND in index_read_map found the next record,
return it.
Otherwise replace the old with a call to index_next (fall through).
*/
if (old_elements != m_queue.elements && part_id != m_top_entry)
{
return_top_record(buf);
DBUG_RETURN(0);
}
}
}
if (part_id >= m_tot_parts)
{
/* This should never happen! */
DBUG_ASSERT(0);
DBUG_RETURN(HA_ERR_END_OF_FILE);
}
file= m_file[part_id];
if (m_index_scan_type == partition_read_range)
{
error= file->read_range_next();
memcpy(rec_buf, table->record[0], m_rec_length);
}
else if (!is_next_same)
error= file->ha_index_next(rec_buf);
else
error= file->ha_index_next_same(rec_buf, m_start_key.key,
m_start_key.length);
if (error)
{
if (error == HA_ERR_END_OF_FILE)
{
/* Return next buffered row */
queue_remove_top(&m_queue);
if (m_queue.elements)
{
DBUG_PRINT("info", ("Record returned from partition %u (2)",
m_top_entry));
return_top_record(buf);
table->status= 0;
error= 0;
}
}
DBUG_RETURN(error);
}
if (!m_using_extended_keys)
{
file->position(rec_buf);
memcpy(rec_buf + m_rec_length, file->ref, file->ref_length);
}
queue_replace_top(&m_queue);
return_top_record(buf);
DBUG_PRINT("info", ("Record returned from partition %u", m_top_entry));
DBUG_RETURN(0);
}
/*
Common routine to handle index_prev with ordered results
SYNOPSIS
handle_ordered_prev()
out:buf Read row in MySQL Row Format
RETURN VALUE
HA_ERR_END_OF_FILE End of scan
0 Success
other Error code
*/
int ha_partition::handle_ordered_prev(uchar *buf)
{
int error;
uint part_id= m_top_entry;
uchar *rec_buf= queue_top(&m_queue) + PARTITION_BYTES_IN_POS;
handler *file= m_file[part_id];
DBUG_ENTER("ha_partition::handle_ordered_prev");
if ((error= file->ha_index_prev(rec_buf)))
{
if (error == HA_ERR_END_OF_FILE)
{
queue_remove_top(&m_queue);
if (m_queue.elements)
{
return_top_record(buf);
DBUG_PRINT("info", ("Record returned from partition %d (2)",
m_top_entry));
error= 0;
table->status= 0;
}
}
DBUG_RETURN(error);
}
queue_replace_top(&m_queue);
return_top_record(buf);
DBUG_PRINT("info", ("Record returned from partition %d", m_top_entry));
DBUG_RETURN(0);
}
/****************************************************************************
MODULE information calls
****************************************************************************/
/*
These are all first approximations of the extra, info, scan_time
and read_time calls
*/
/**
Helper function for sorting according to number of rows in descending order.
*/
int ha_partition::compare_number_of_records(ha_partition *me,
const uint32 *a,
const uint32 *b)
{
handler **file= me->m_file;
/* Note: sorting in descending order! */
if (file[*a]->stats.records > file[*b]->stats.records)
return -1;
if (file[*a]->stats.records < file[*b]->stats.records)
return 1;
return 0;
}
/*
General method to gather info from handler
SYNOPSIS
info()
flag Specifies what info is requested
RETURN VALUE
NONE
DESCRIPTION
::info() is used to return information to the optimizer.
Currently this table handler doesn't implement most of the fields
really needed. SHOW also makes use of this data
Another note, if your handler doesn't proved exact record count,
you will probably want to have the following in your code:
if (records < 2)
records = 2;
The reason is that the server will optimize for cases of only a single
record. If in a table scan you don't know the number of records
it will probably be better to set records to two so you can return
as many records as you need.
Along with records a few more variables you may wish to set are:
records
deleted
data_file_length
index_file_length
delete_length
check_time
Take a look at the public variables in handler.h for more information.
Called in:
filesort.cc
ha_heap.cc
item_sum.cc
opt_sum.cc
sql_delete.cc
sql_delete.cc
sql_derived.cc
sql_select.cc
sql_select.cc
sql_select.cc
sql_select.cc
sql_select.cc
sql_show.cc
sql_show.cc
sql_show.cc
sql_show.cc
sql_table.cc
sql_union.cc
sql_update.cc
Some flags that are not implemented
HA_STATUS_POS:
This parameter is never used from the MySQL Server. It is checked in a
place in MyISAM so could potentially be used by MyISAM specific
programs.
HA_STATUS_NO_LOCK:
This is declared and often used. It's only used by MyISAM.
It means that MySQL doesn't need the absolute latest statistics
information. This may save the handler from doing internal locks while
retrieving statistics data.
*/
int ha_partition::info(uint flag)
{
uint no_lock_flag= flag & HA_STATUS_NO_LOCK;
uint extra_var_flag= flag & HA_STATUS_VARIABLE_EXTRA;
DBUG_ENTER("ha_partition::info");
#ifndef DBUG_OFF
if (bitmap_is_set_all(&(m_part_info->read_partitions)))
DBUG_PRINT("info", ("All partitions are used"));
#endif /* DBUG_OFF */
if (flag & HA_STATUS_AUTO)
{
bool auto_inc_is_first_in_idx= (table_share->next_number_keypart == 0);
DBUG_PRINT("info", ("HA_STATUS_AUTO"));
if (!table->found_next_number_field)
stats.auto_increment_value= 0;
else if (part_share->auto_inc_initialized)
{
lock_auto_increment();
stats.auto_increment_value= part_share->next_auto_inc_val;
unlock_auto_increment();
}
else
{
lock_auto_increment();
/* to avoid two concurrent initializations, check again when locked */
if (part_share->auto_inc_initialized)
stats.auto_increment_value= part_share->next_auto_inc_val;
else
{
/*
The auto-inc mutex in the table_share is locked, so we do not need
to have the handlers locked.
HA_STATUS_NO_LOCK is not checked, since we cannot skip locking
the mutex, because it is initialized.
*/
handler *file, **file_array;
ulonglong auto_increment_value= 0;
file_array= m_file;
DBUG_PRINT("info",
("checking all partitions for auto_increment_value"));
do
{
file= *file_array;
file->info(HA_STATUS_AUTO | no_lock_flag);
set_if_bigger(auto_increment_value,
file->stats.auto_increment_value);
} while (*(++file_array));
DBUG_ASSERT(auto_increment_value);
stats.auto_increment_value= auto_increment_value;
if (auto_inc_is_first_in_idx)
{
set_if_bigger(part_share->next_auto_inc_val,
auto_increment_value);
part_share->auto_inc_initialized= true;
DBUG_PRINT("info", ("initializing next_auto_inc_val to %lu",
(ulong) part_share->next_auto_inc_val));
}
}
unlock_auto_increment();
}
}
if (flag & HA_STATUS_VARIABLE)
{
uint i;
DBUG_PRINT("info", ("HA_STATUS_VARIABLE"));
/*
Calculates statistical variables
records: Estimate of number records in table
We report sum (always at least 2 if not empty)
deleted: Estimate of number holes in the table due to
deletes
We report sum
data_file_length: Length of data file, in principle bytes in table
We report sum
index_file_length: Length of index file, in principle bytes in
indexes in the table
We report sum
delete_length: Length of free space easily used by new records in table
We report sum
mean_record_length:Mean record length in the table
We calculate this
check_time: Time of last check (only applicable to MyISAM)
We report last time of all underlying handlers
*/
handler *file;
stats.records= 0;
stats.deleted= 0;
stats.data_file_length= 0;
stats.index_file_length= 0;
stats.check_time= 0;
stats.delete_length= 0;
for (i= bitmap_get_first_set(&m_part_info->read_partitions);
i < m_tot_parts;
i= bitmap_get_next_set(&m_part_info->read_partitions, i))
{
file= m_file[i];
file->info(HA_STATUS_VARIABLE | no_lock_flag | extra_var_flag);
stats.records+= file->stats.records;
stats.deleted+= file->stats.deleted;
stats.data_file_length+= file->stats.data_file_length;
stats.index_file_length+= file->stats.index_file_length;
stats.delete_length+= file->stats.delete_length;
if (file->stats.check_time > stats.check_time)
stats.check_time= file->stats.check_time;
}
if (stats.records && stats.records < 2 &&
!(m_file[0]->ha_table_flags() & HA_STATS_RECORDS_IS_EXACT))
stats.records= 2;
if (stats.records > 0)
stats.mean_rec_length= (ulong) (stats.data_file_length / stats.records);
else
stats.mean_rec_length= 0;
}
if (flag & HA_STATUS_CONST)
{
DBUG_PRINT("info", ("HA_STATUS_CONST"));
/*
Recalculate loads of constant variables. MyISAM also sets things
directly on the table share object.
Check whether this should be fixed since handlers should not
change things directly on the table object.
Monty comment: This should NOT be changed! It's the handlers
responsibility to correct table->s->keys_xxxx information if keys
have been disabled.
The most important parameters set here is records per key on
all indexes. block_size and primar key ref_length.
For each index there is an array of rec_per_key.
As an example if we have an index with three attributes a,b and c
we will have an array of 3 rec_per_key.
rec_per_key[0] is an estimate of number of records divided by
number of unique values of the field a.
rec_per_key[1] is an estimate of the number of records divided
by the number of unique combinations of the fields a and b.
rec_per_key[2] is an estimate of the number of records divided
by the number of unique combinations of the fields a,b and c.
Many handlers only set the value of rec_per_key when all fields
are bound (rec_per_key[2] in the example above).
If the handler doesn't support statistics, it should set all of the
above to 0.
We first scans through all partitions to get the one holding most rows.
We will then allow the handler with the most rows to set
the rec_per_key and use this as an estimate on the total table.
max_data_file_length: Maximum data file length
We ignore it, is only used in
SHOW TABLE STATUS
max_index_file_length: Maximum index file length
We ignore it since it is never used
block_size: Block size used
We set it to the value of the first handler
ref_length: We set this to the value calculated
and stored in local object
create_time: Creation time of table
So we calculate these constants by using the variables from the
handler with most rows.
*/
handler *file, **file_array;
ulonglong max_records= 0;
uint32 i= 0;
uint32 handler_instance= 0;
file_array= m_file;
do
{
file= *file_array;
/* Get variables if not already done */
if (!(flag & HA_STATUS_VARIABLE) ||
!bitmap_is_set(&(m_part_info->read_partitions),
(uint)(file_array - m_file)))
file->info(HA_STATUS_VARIABLE | no_lock_flag | extra_var_flag);
if (file->stats.records > max_records)
{
max_records= file->stats.records;
handler_instance= i;
}
i++;
} while (*(++file_array));
/*
Sort the array of part_ids by number of records in
in descending order.
*/
my_qsort2((void*) m_part_ids_sorted_by_num_of_records,
m_tot_parts,
sizeof(uint32),
(qsort2_cmp) compare_number_of_records,
this);
file= m_file[handler_instance];
file->info(HA_STATUS_CONST | no_lock_flag);
stats.block_size= file->stats.block_size;
stats.create_time= file->stats.create_time;
ref_length= m_ref_length;
}
if (flag & HA_STATUS_ERRKEY)
{
handler *file= m_file[m_last_part];
DBUG_PRINT("info", ("info: HA_STATUS_ERRKEY"));
/*
This flag is used to get index number of the unique index that
reported duplicate key
We will report the errkey on the last handler used and ignore the rest
Note: all engines does not support HA_STATUS_ERRKEY, so set errkey.
*/
file->errkey= errkey;
file->info(HA_STATUS_ERRKEY | no_lock_flag);
errkey= file->errkey;
}
if (flag & HA_STATUS_TIME)
{
handler *file, **file_array;
DBUG_PRINT("info", ("info: HA_STATUS_TIME"));
/*
This flag is used to set the latest update time of the table.
Used by SHOW commands
We will report the maximum of these times
*/
stats.update_time= 0;
file_array= m_file;
do
{
file= *file_array;
file->info(HA_STATUS_TIME | no_lock_flag);
if (file->stats.update_time > stats.update_time)
stats.update_time= file->stats.update_time;
} while (*(++file_array));
}
DBUG_RETURN(0);
}
void ha_partition::get_dynamic_partition_info(PARTITION_STATS *stat_info,
uint part_id)
{
handler *file= m_file[part_id];
DBUG_ASSERT(bitmap_is_set(&(m_part_info->read_partitions), part_id));
file->info(HA_STATUS_TIME | HA_STATUS_VARIABLE |
HA_STATUS_VARIABLE_EXTRA | HA_STATUS_NO_LOCK);
stat_info->records= file->stats.records;
stat_info->mean_rec_length= file->stats.mean_rec_length;
stat_info->data_file_length= file->stats.data_file_length;
stat_info->max_data_file_length= file->stats.max_data_file_length;
stat_info->index_file_length= file->stats.index_file_length;
stat_info->delete_length= file->stats.delete_length;
stat_info->create_time= file->stats.create_time;
stat_info->update_time= file->stats.update_time;
stat_info->check_time= file->stats.check_time;
stat_info->check_sum= 0;
if (file->ha_table_flags() & (HA_HAS_OLD_CHECKSUM | HA_HAS_NEW_CHECKSUM))
stat_info->check_sum= file->checksum();
return;
}
/**
General function to prepare handler for certain behavior.
@param[in] operation operation to execute
@return status
@retval 0 success
@retval >0 error code
@detail
extra() is called whenever the server wishes to send a hint to
the storage engine. The MyISAM engine implements the most hints.
We divide the parameters into the following categories:
1) Operations used by most handlers
2) Operations used by some non-MyISAM handlers
3) Operations used only by MyISAM
4) Operations only used by temporary tables for query processing
5) Operations only used by MyISAM internally
6) Operations not used at all
7) Operations only used by federated tables for query processing
8) Operations only used by NDB
9) Operations only used by MERGE
The partition handler need to handle category 1), 2) and 3).
1) Operations used by most handlers
-----------------------------------
HA_EXTRA_RESET:
This option is used by most handlers and it resets the handler state
to the same state as after an open call. This includes releasing
any READ CACHE or WRITE CACHE or other internal buffer used.
It is called from the reset method in the handler interface. There are
three instances where this is called.
1) After completing a INSERT ... SELECT ... query the handler for the
table inserted into is reset
2) It is called from close_thread_table which in turn is called from
close_thread_tables except in the case where the tables are locked
in which case ha_commit_stmt is called instead.
It is only called from here if refresh_version hasn't changed and the
table is not an old table when calling close_thread_table.
close_thread_tables is called from many places as a general clean up
function after completing a query.
3) It is called when deleting the QUICK_RANGE_SELECT object if the
QUICK_RANGE_SELECT object had its own handler object. It is called
immediatley before close of this local handler object.
HA_EXTRA_KEYREAD:
HA_EXTRA_NO_KEYREAD:
These parameters are used to provide an optimisation hint to the handler.
If HA_EXTRA_KEYREAD is set it is enough to read the index fields, for
many handlers this means that the index-only scans can be used and it
is not necessary to use the real records to satisfy this part of the
query. Index-only scans is a very important optimisation for disk-based
indexes. For main-memory indexes most indexes contain a reference to the
record and thus KEYREAD only says that it is enough to read key fields.
HA_EXTRA_NO_KEYREAD disables this for the handler, also HA_EXTRA_RESET
will disable this option.
The handler will set HA_KEYREAD_ONLY in its table flags to indicate this
feature is supported.
HA_EXTRA_FLUSH:
Indication to flush tables to disk, is supposed to be used to
ensure disk based tables are flushed at end of query execution.
Currently is never used.
HA_EXTRA_FORCE_REOPEN:
Only used by MyISAM and Archive, called when altering table,
closing tables to enforce a reopen of the table files.
2) Operations used by some non-MyISAM handlers
----------------------------------------------
HA_EXTRA_KEYREAD_PRESERVE_FIELDS:
This is a strictly InnoDB feature that is more or less undocumented.
When it is activated InnoDB copies field by field from its fetch
cache instead of all fields in one memcpy. Have no idea what the
purpose of this is.
Cut from include/my_base.h:
When using HA_EXTRA_KEYREAD, overwrite only key member fields and keep
other fields intact. When this is off (by default) InnoDB will use memcpy
to overwrite entire row.
HA_EXTRA_IGNORE_DUP_KEY:
HA_EXTRA_NO_IGNORE_DUP_KEY:
Informs the handler to we will not stop the transaction if we get an
duplicate key errors during insert/upate.
Always called in pair, triggered by INSERT IGNORE and other similar
SQL constructs.
Not used by MyISAM.
3) Operations used only by MyISAM
---------------------------------
HA_EXTRA_NORMAL:
Only used in MyISAM to reset quick mode, not implemented by any other
handler. Quick mode is also reset in MyISAM by HA_EXTRA_RESET.
It is called after completing a successful DELETE query if the QUICK
option is set.
HA_EXTRA_QUICK:
When the user does DELETE QUICK FROM table where-clause; this extra
option is called before the delete query is performed and
HA_EXTRA_NORMAL is called after the delete query is completed.
Temporary tables used internally in MySQL always set this option
The meaning of quick mode is that when deleting in a B-tree no merging
of leafs is performed. This is a common method and many large DBMS's
actually only support this quick mode since it is very difficult to
merge leaves in a tree used by many threads concurrently.
HA_EXTRA_CACHE:
This flag is usually set with extra_opt along with a cache size.
The size of this buffer is set by the user variable
record_buffer_size. The value of this cache size is the amount of
data read from disk in each fetch when performing a table scan.
This means that before scanning a table it is normal to call
extra with HA_EXTRA_CACHE and when the scan is completed to call
HA_EXTRA_NO_CACHE to release the cache memory.
Some special care is taken when using this extra parameter since there
could be a write ongoing on the table in the same statement. In this
one has to take special care since there might be a WRITE CACHE as
well. HA_EXTRA_CACHE specifies using a READ CACHE and using
READ CACHE and WRITE CACHE at the same time is not possible.
Only MyISAM currently use this option.
It is set when doing full table scans using rr_sequential and
reset when completing such a scan with end_read_record
(resetting means calling extra with HA_EXTRA_NO_CACHE).
It is set in filesort.cc for MyISAM internal tables and it is set in
a multi-update where HA_EXTRA_CACHE is called on a temporary result
table and after that ha_rnd_init(0) on table to be updated
and immediately after that HA_EXTRA_NO_CACHE on table to be updated.
Apart from that it is always used from init_read_record but not when
used from UPDATE statements. It is not used from DELETE statements
with ORDER BY and LIMIT but it is used in normal scan loop in DELETE
statements. The reason here is that DELETE's in MyISAM doesn't move
existings data rows.
It is also set in copy_data_between_tables when scanning the old table
to copy over to the new table.
And it is set in join_init_read_record where quick objects are used
to perform a scan on the table. In this case the full table scan can
even be performed multiple times as part of the nested loop join.
For purposes of the partition handler it is obviously necessary to have
special treatment of this extra call. If we would simply pass this
extra call down to each handler we would allocate
cache size * no of partitions amount of memory and this is not
necessary since we will only scan one partition at a time when doing
full table scans.
Thus we treat it by first checking whether we have MyISAM handlers in
the table, if not we simply ignore the call and if we have we will
record the call but will not call any underlying handler yet. Then
when performing the sequential scan we will check this recorded value
and call extra_opt whenever we start scanning a new partition.
HA_EXTRA_NO_CACHE:
When performing a UNION SELECT HA_EXTRA_NO_CACHE is called from the
flush method in the select_union class.
It is used to some extent when insert delayed inserts.
See HA_EXTRA_RESET_STATE for use in conjunction with delete_all_rows().
It should be ok to call HA_EXTRA_NO_CACHE on all underlying handlers
if they are MyISAM handlers. Other handlers we can ignore the call
for. If no cache is in use they will quickly return after finding
this out. And we also ensure that all caches are disabled and no one
is left by mistake.
In the future this call will probably be deleted and we will instead call
::reset();
HA_EXTRA_WRITE_CACHE:
See above, called from various places. It is mostly used when we
do INSERT ... SELECT
No special handling to save cache space is developed currently.
HA_EXTRA_PREPARE_FOR_UPDATE:
This is called as part of a multi-table update. When the table to be
updated is also scanned then this informs MyISAM handler to drop any
caches if dynamic records are used (fixed size records do not care
about this call). We pass this along to the first partition to scan, and
flag that it is to be called after HA_EXTRA_CACHE when moving to the next
partition to scan.
HA_EXTRA_PREPARE_FOR_DROP:
Only used by MyISAM, called in preparation for a DROP TABLE.
It's used mostly by Windows that cannot handle dropping an open file.
On other platforms it has the same effect as HA_EXTRA_FORCE_REOPEN.
HA_EXTRA_PREPARE_FOR_RENAME:
Informs the handler we are about to attempt a rename of the table.
For handlers that have share open files (MyISAM key-file and
Archive writer) they must close the files before rename is possible
on Windows.
HA_EXTRA_READCHECK:
HA_EXTRA_NO_READCHECK:
Only one call to HA_EXTRA_NO_READCHECK from ha_open where it says that
this is not needed in SQL. The reason for this call is that MyISAM sets
the READ_CHECK_USED in the open call so the call is needed for MyISAM
to reset this feature.
The idea with this parameter was to inform of doing/not doing a read
check before applying an update. Since SQL always performs a read before
applying the update No Read Check is needed in MyISAM as well.
This is a cut from Docs/myisam.txt
Sometimes you might want to force an update without checking whether
another user has changed the record since you last read it. This is
somewhat dangerous, so it should ideally not be used. That can be
accomplished by wrapping the mi_update() call in two calls to mi_extra(),
using these functions:
HA_EXTRA_NO_READCHECK=5 No readcheck on update
HA_EXTRA_READCHECK=6 Use readcheck (def)
4) Operations only used by temporary tables for query processing
----------------------------------------------------------------
HA_EXTRA_RESET_STATE:
Same as reset() except that buffers are not released. If there is
a READ CACHE it is reinit'ed. A cache is reinit'ed to restart reading
or to change type of cache between READ CACHE and WRITE CACHE.
This extra function is always called immediately before calling
delete_all_rows on the handler for temporary tables.
There are cases however when HA_EXTRA_RESET_STATE isn't called in
a similar case for a temporary table in sql_union.cc and in two other
cases HA_EXTRA_NO_CACHE is called before and HA_EXTRA_WRITE_CACHE
called afterwards.
The case with HA_EXTRA_NO_CACHE and HA_EXTRA_WRITE_CACHE means
disable caching, delete all rows and enable WRITE CACHE. This is
used for temporary tables containing distinct sums and a
functional group.
The only case that delete_all_rows is called on non-temporary tables
is in sql_delete.cc when DELETE FROM table; is called by a user.
In this case no special extra calls are performed before or after this
call.
The partition handler should not need to bother about this one. It
should never be called.
HA_EXTRA_NO_ROWS:
Don't insert rows indication to HEAP and MyISAM, only used by temporary
tables used in query processing.
Not handled by partition handler.
5) Operations only used by MyISAM internally
--------------------------------------------
HA_EXTRA_REINIT_CACHE:
This call reinitializes the READ CACHE described above if there is one
and otherwise the call is ignored.
We can thus safely call it on all underlying handlers if they are
MyISAM handlers. It is however never called so we don't handle it at all.
HA_EXTRA_FLUSH_CACHE:
Flush WRITE CACHE in MyISAM. It is only from one place in the code.
This is in sql_insert.cc where it is called if the table_flags doesn't
contain HA_DUPLICATE_POS. The only handler having the HA_DUPLICATE_POS
set is the MyISAM handler and so the only handler not receiving this
call is MyISAM.
Thus in effect this call is called but never used. Could be removed
from sql_insert.cc
HA_EXTRA_NO_USER_CHANGE:
Only used by MyISAM, never called.
Simulates lock_type as locked.
HA_EXTRA_WAIT_LOCK:
HA_EXTRA_WAIT_NOLOCK:
Only used by MyISAM, called from MyISAM handler but never from server
code on top of the handler.
Sets lock_wait on/off
HA_EXTRA_NO_KEYS:
Only used MyISAM, only used internally in MyISAM handler, never called
from server level.
HA_EXTRA_KEYREAD_CHANGE_POS:
HA_EXTRA_REMEMBER_POS:
HA_EXTRA_RESTORE_POS:
HA_EXTRA_PRELOAD_BUFFER_SIZE:
HA_EXTRA_CHANGE_KEY_TO_DUP:
HA_EXTRA_CHANGE_KEY_TO_UNIQUE:
Only used by MyISAM, never called.
6) Operations not used at all
-----------------------------
HA_EXTRA_KEY_CACHE:
HA_EXTRA_NO_KEY_CACHE:
This parameters are no longer used and could be removed.
7) Operations only used by federated tables for query processing
----------------------------------------------------------------
HA_EXTRA_INSERT_WITH_UPDATE:
Inform handler that an "INSERT...ON DUPLICATE KEY UPDATE" will be
executed. This condition is unset by HA_EXTRA_NO_IGNORE_DUP_KEY.
8) Operations only used by NDB
------------------------------
HA_EXTRA_DELETE_CANNOT_BATCH:
HA_EXTRA_UPDATE_CANNOT_BATCH:
Inform handler that delete_row()/update_row() cannot batch deletes/updates
and should perform them immediately. This may be needed when table has
AFTER DELETE/UPDATE triggers which access to subject table.
These flags are reset by the handler::extra(HA_EXTRA_RESET) call.
9) Operations only used by MERGE
------------------------------
HA_EXTRA_ADD_CHILDREN_LIST:
HA_EXTRA_ATTACH_CHILDREN:
HA_EXTRA_IS_ATTACHED_CHILDREN:
HA_EXTRA_DETACH_CHILDREN:
Special actions for MERGE tables. Ignore.
*/
int ha_partition::extra(enum ha_extra_function operation)
{
DBUG_ENTER("ha_partition:extra");
DBUG_PRINT("info", ("operation: %d", (int) operation));
switch (operation) {
/* Category 1), used by most handlers */
case HA_EXTRA_KEYREAD:
case HA_EXTRA_NO_KEYREAD:
case HA_EXTRA_FLUSH:
case HA_EXTRA_PREPARE_FOR_FORCED_CLOSE:
DBUG_RETURN(loop_extra(operation));
case HA_EXTRA_PREPARE_FOR_RENAME:
case HA_EXTRA_FORCE_REOPEN:
DBUG_RETURN(loop_extra_alter(operation));
break;
/* Category 2), used by non-MyISAM handlers */
case HA_EXTRA_IGNORE_DUP_KEY:
case HA_EXTRA_NO_IGNORE_DUP_KEY:
case HA_EXTRA_KEYREAD_PRESERVE_FIELDS:
{
if (!m_myisam)
DBUG_RETURN(loop_extra(operation));
break;
}
/* Category 3), used by MyISAM handlers */
case HA_EXTRA_PREPARE_FOR_UPDATE:
/*
Needs to be run on the first partition in the range now, and
later in late_extra_cache, when switching to a new partition to scan.
*/
m_extra_prepare_for_update= TRUE;
if (m_part_spec.start_part != NO_CURRENT_PART_ID)
{
if (!m_extra_cache)
m_extra_cache_part_id= m_part_spec.start_part;
DBUG_ASSERT(m_extra_cache_part_id == m_part_spec.start_part);
(void) m_file[m_part_spec.start_part]->extra(HA_EXTRA_PREPARE_FOR_UPDATE);
}
break;
case HA_EXTRA_NORMAL:
case HA_EXTRA_QUICK:
case HA_EXTRA_PREPARE_FOR_DROP:
case HA_EXTRA_FLUSH_CACHE:
case HA_EXTRA_PREPARE_FOR_ALTER_TABLE:
{
DBUG_RETURN(loop_extra(operation));
}
case HA_EXTRA_NO_READCHECK:
{
/*
This is only done as a part of ha_open, which is also used in
ha_partition::open, so no need to do anything.
*/
break;
}
case HA_EXTRA_CACHE:
{
prepare_extra_cache(0);
break;
}
case HA_EXTRA_NO_CACHE:
{
int ret= 0;
if (m_extra_cache_part_id != NO_CURRENT_PART_ID)
ret= m_file[m_extra_cache_part_id]->extra(HA_EXTRA_NO_CACHE);
m_extra_cache= FALSE;
m_extra_cache_size= 0;
m_extra_prepare_for_update= FALSE;
m_extra_cache_part_id= NO_CURRENT_PART_ID;
DBUG_RETURN(ret);
}
case HA_EXTRA_WRITE_CACHE:
{
m_extra_cache= FALSE;
m_extra_cache_size= 0;
m_extra_prepare_for_update= FALSE;
m_extra_cache_part_id= NO_CURRENT_PART_ID;
DBUG_RETURN(loop_extra(operation));
}
case HA_EXTRA_IGNORE_NO_KEY:
case HA_EXTRA_NO_IGNORE_NO_KEY:
{
/*
Ignore as these are specific to NDB for handling
idempotency
*/
break;
}
case HA_EXTRA_WRITE_CAN_REPLACE:
case HA_EXTRA_WRITE_CANNOT_REPLACE:
{
/*
Informs handler that write_row() can replace rows which conflict
with row being inserted by PK/unique key without reporting error
to the SQL-layer.
This optimization is not safe for partitioned table in general case
since we may have to put new version of row into partition which is
different from partition in which old version resides (for example
when we partition by non-PK column or by some column which is not
part of unique key which were violated).
And since NDB which is the only engine at the moment that supports
this optimization handles partitioning on its own we simple disable
it here. (BTW for NDB this optimization is safe since it supports
only KEY partitioning and won't use this optimization for tables
which have additional unique constraints).
*/
break;
}
/* Category 7), used by federated handlers */
case HA_EXTRA_INSERT_WITH_UPDATE:
DBUG_RETURN(loop_extra(operation));
/* Category 8) Operations only used by NDB */
case HA_EXTRA_DELETE_CANNOT_BATCH:
case HA_EXTRA_UPDATE_CANNOT_BATCH:
{
/* Currently only NDB use the *_CANNOT_BATCH */
break;
}
/* Category 9) Operations only used by MERGE */
case HA_EXTRA_ADD_CHILDREN_LIST:
case HA_EXTRA_ATTACH_CHILDREN:
case HA_EXTRA_IS_ATTACHED_CHILDREN:
case HA_EXTRA_DETACH_CHILDREN:
{
/* Special actions for MERGE tables. Ignore. */
break;
}
/*
http://dev.mysql.com/doc/refman/5.1/en/partitioning-limitations.html
says we no longer support logging to partitioned tables, so we fail
here.
*/
case HA_EXTRA_MARK_AS_LOG_TABLE:
DBUG_RETURN(ER_UNSUPORTED_LOG_ENGINE);
default:
{
/* Temporary crash to discover what is wrong */
DBUG_ASSERT(0);
break;
}
}
DBUG_RETURN(0);
}
/**
Special extra call to reset extra parameters
@return Operation status.
@retval >0 Error code
@retval 0 Success
@note Called at end of each statement to reset buffers.
To avoid excessive calls, the m_partitions_to_reset bitmap keep records
of which partitions that have been used in extra(), external_lock() or
start_stmt() and is needed to be called.
*/
int ha_partition::reset(void)
{
int result= 0;
int tmp;
uint i;
DBUG_ENTER("ha_partition::reset");
for (i= bitmap_get_first_set(&m_partitions_to_reset);
i < m_tot_parts;
i= bitmap_get_next_set(&m_partitions_to_reset, i))
{
if ((tmp= m_file[i]->ha_reset()))
result= tmp;
}
bitmap_clear_all(&m_partitions_to_reset);
m_extra_prepare_for_update= FALSE;
DBUG_RETURN(result);
}
/*
Special extra method for HA_EXTRA_CACHE with cachesize as extra parameter
SYNOPSIS
extra_opt()
operation Must be HA_EXTRA_CACHE
cachesize Size of cache in full table scan
RETURN VALUE
>0 Error code
0 Success
*/
int ha_partition::extra_opt(enum ha_extra_function operation, ulong cachesize)
{
DBUG_ENTER("ha_partition::extra_opt()");
DBUG_ASSERT(HA_EXTRA_CACHE == operation);
prepare_extra_cache(cachesize);
DBUG_RETURN(0);
}
/*
Call extra on handler with HA_EXTRA_CACHE and cachesize
SYNOPSIS
prepare_extra_cache()
cachesize Size of cache for full table scan
RETURN VALUE
NONE
*/
void ha_partition::prepare_extra_cache(uint cachesize)
{
DBUG_ENTER("ha_partition::prepare_extra_cache()");
DBUG_PRINT("info", ("cachesize %u", cachesize));
m_extra_cache= TRUE;
m_extra_cache_size= cachesize;
if (m_part_spec.start_part != NO_CURRENT_PART_ID)
{
DBUG_ASSERT(bitmap_is_set(&m_partitions_to_reset,
m_part_spec.start_part));
bitmap_set_bit(&m_partitions_to_reset, m_part_spec.start_part);
late_extra_cache(m_part_spec.start_part);
}
DBUG_VOID_RETURN;
}
/**
Prepares our new and reorged handlers for rename or delete.
@param operation Operation to forward
@return Operation status
@retval 0 Success
@retval !0 Error
*/
int ha_partition::loop_extra_alter(enum ha_extra_function operation)
{
int result= 0, tmp;
handler **file;
DBUG_ENTER("ha_partition::loop_extra_alter()");
DBUG_ASSERT(operation == HA_EXTRA_PREPARE_FOR_RENAME ||
operation == HA_EXTRA_FORCE_REOPEN);
if (m_new_file != NULL)
{
for (file= m_new_file; *file; file++)
if ((tmp= (*file)->extra(operation)))
result= tmp;
}
if (m_reorged_file != NULL)
{
for (file= m_reorged_file; *file; file++)
if ((tmp= (*file)->extra(operation)))
result= tmp;
}
if ((tmp= loop_extra(operation)))
result= tmp;
DBUG_RETURN(result);
}
/*
Call extra on all partitions
SYNOPSIS
loop_extra()
operation extra operation type
RETURN VALUE
>0 Error code
0 Success
*/
int ha_partition::loop_extra(enum ha_extra_function operation)
{
int result= 0, tmp;
uint i;
DBUG_ENTER("ha_partition::loop_extra()");
for (i= bitmap_get_first_set(&m_part_info->lock_partitions);
i < m_tot_parts;
i= bitmap_get_next_set(&m_part_info->lock_partitions, i))
{
if ((tmp= m_file[i]->extra(operation)))
result= tmp;
}
/* Add all used partitions to be called in reset(). */
bitmap_union(&m_partitions_to_reset, &m_part_info->lock_partitions);
DBUG_RETURN(result);
}
/*
Call extra(HA_EXTRA_CACHE) on next partition_id
SYNOPSIS
late_extra_cache()
partition_id Partition id to call extra on
RETURN VALUE
NONE
*/
void ha_partition::late_extra_cache(uint partition_id)
{
handler *file;
DBUG_ENTER("ha_partition::late_extra_cache");
DBUG_PRINT("info", ("extra_cache %u prepare %u partid %u size %u",
m_extra_cache, m_extra_prepare_for_update,
partition_id, m_extra_cache_size));
if (!m_extra_cache && !m_extra_prepare_for_update)
DBUG_VOID_RETURN;
file= m_file[partition_id];
if (m_extra_cache)
{
if (m_extra_cache_size == 0)
(void) file->extra(HA_EXTRA_CACHE);
else
(void) file->extra_opt(HA_EXTRA_CACHE, m_extra_cache_size);
}
if (m_extra_prepare_for_update)
{
DBUG_ASSERT(m_extra_cache);
(void) file->extra(HA_EXTRA_PREPARE_FOR_UPDATE);
}
m_extra_cache_part_id= partition_id;
DBUG_VOID_RETURN;
}
/*
Call extra(HA_EXTRA_NO_CACHE) on next partition_id
SYNOPSIS
late_extra_no_cache()
partition_id Partition id to call extra on
RETURN VALUE
NONE
*/
void ha_partition::late_extra_no_cache(uint partition_id)
{
handler *file;
DBUG_ENTER("ha_partition::late_extra_no_cache");
if (!m_extra_cache && !m_extra_prepare_for_update)
DBUG_VOID_RETURN;
file= m_file[partition_id];
(void) file->extra(HA_EXTRA_NO_CACHE);
DBUG_ASSERT(partition_id == m_extra_cache_part_id);
m_extra_cache_part_id= NO_CURRENT_PART_ID;
DBUG_VOID_RETURN;
}
/****************************************************************************
MODULE optimiser support
****************************************************************************/
/**
Get keys to use for scanning.
@return key_map of keys usable for scanning
@note No need to use read_partitions here, since it does not depend on
which partitions is used, only which storage engine used.
*/
const key_map *ha_partition::keys_to_use_for_scanning()
{
DBUG_ENTER("ha_partition::keys_to_use_for_scanning");
DBUG_RETURN(m_file[0]->keys_to_use_for_scanning());
}
/**
Minimum number of rows to base optimizer estimate on.
*/
ha_rows ha_partition::min_rows_for_estimate()
{
uint i, max_used_partitions, tot_used_partitions;
DBUG_ENTER("ha_partition::min_rows_for_estimate");
tot_used_partitions= bitmap_bits_set(&m_part_info->read_partitions);
/*
All partitions might have been left as unused during partition pruning
due to, for example, an impossible WHERE condition. Nonetheless, the
optimizer might still attempt to perform (e.g. range) analysis where an
estimate of the the number of rows is calculated using records_in_range.
Hence, to handle this and other possible cases, use zero as the minimum
number of rows to base the estimate on if no partition is being used.
*/
if (!tot_used_partitions)
DBUG_RETURN(0);
/*
Allow O(log2(tot_partitions)) increase in number of used partitions.
This gives O(tot_rows/log2(tot_partitions)) rows to base the estimate on.
I.e when the total number of partitions doubles, allow one more
partition to be checked.
*/
i= 2;
max_used_partitions= 1;
while (i < m_tot_parts)
{
max_used_partitions++;
i= i << 1;
}
if (max_used_partitions > tot_used_partitions)
max_used_partitions= tot_used_partitions;
/* stats.records is already updated by the info(HA_STATUS_VARIABLE) call. */
DBUG_PRINT("info", ("max_used_partitions: %u tot_rows: %lu",
max_used_partitions,
(ulong) stats.records));
DBUG_PRINT("info", ("tot_used_partitions: %u min_rows_to_check: %lu",
tot_used_partitions,
(ulong) stats.records * max_used_partitions
/ tot_used_partitions));
DBUG_RETURN(stats.records * max_used_partitions / tot_used_partitions);
}
/**
Get the biggest used partition.
Starting at the N:th biggest partition and skips all non used
partitions, returning the biggest used partition found
@param[in,out] part_index Skip the *part_index biggest partitions
@return The biggest used partition with index not lower than *part_index.
@retval NO_CURRENT_PART_ID No more partition used.
@retval != NO_CURRENT_PART_ID partition id of biggest used partition with
index >= *part_index supplied. Note that
*part_index will be updated to the next
partition index to use.
*/
uint ha_partition::get_biggest_used_partition(uint *part_index)
{
uint part_id;
while ((*part_index) < m_tot_parts)
{
part_id= m_part_ids_sorted_by_num_of_records[(*part_index)++];
if (bitmap_is_set(&m_part_info->read_partitions, part_id))
return part_id;
}
return NO_CURRENT_PART_ID;
}
/*
Return time for a scan of the table
SYNOPSIS
scan_time()
RETURN VALUE
time for scan
*/
double ha_partition::scan_time()
{
double scan_time= 0;
uint i;
DBUG_ENTER("ha_partition::scan_time");
for (i= bitmap_get_first_set(&m_part_info->read_partitions);
i < m_tot_parts;
i= bitmap_get_next_set(&m_part_info->read_partitions, i))
scan_time+= m_file[i]->scan_time();
DBUG_RETURN(scan_time);
}
/**
Find number of records in a range.
@param inx Index number
@param min_key Start of range
@param max_key End of range
@return Number of rows in range.
Given a starting key, and an ending key estimate the number of rows that
will exist between the two. max_key may be empty which in case determine
if start_key matches any rows.
*/
ha_rows ha_partition::records_in_range(uint inx, key_range *min_key,
key_range *max_key)
{
ha_rows min_rows_to_check, rows, estimated_rows=0, checked_rows= 0;
uint partition_index= 0, part_id;
DBUG_ENTER("ha_partition::records_in_range");
min_rows_to_check= min_rows_for_estimate();
while ((part_id= get_biggest_used_partition(&partition_index))
!= NO_CURRENT_PART_ID)
{
rows= m_file[part_id]->records_in_range(inx, min_key, max_key);
DBUG_PRINT("info", ("part %u match %lu rows of %lu", part_id, (ulong) rows,
(ulong) m_file[part_id]->stats.records));
if (rows == HA_POS_ERROR)
DBUG_RETURN(HA_POS_ERROR);
estimated_rows+= rows;
checked_rows+= m_file[part_id]->stats.records;
/*
Returning 0 means no rows can be found, so we must continue
this loop as long as we have estimated_rows == 0.
Also many engines return 1 to indicate that there may exist
a matching row, we do not normalize this by dividing by number of
used partitions, but leave it to be returned as a sum, which will
reflect that we will need to scan each partition's index.
Note that this statistics may not always be correct, so we must
continue even if the current partition has 0 rows, since we might have
deleted rows from the current partition, or inserted to the next
partition.
*/
if (estimated_rows && checked_rows &&
checked_rows >= min_rows_to_check)
{
DBUG_PRINT("info",
("records_in_range(inx %u): %lu (%lu * %lu / %lu)",
inx,
(ulong) (estimated_rows * stats.records / checked_rows),
(ulong) estimated_rows,
(ulong) stats.records,
(ulong) checked_rows));
DBUG_RETURN(estimated_rows * stats.records / checked_rows);
}
}
DBUG_PRINT("info", ("records_in_range(inx %u): %lu",
inx,
(ulong) estimated_rows));
DBUG_RETURN(estimated_rows);
}
/**
Estimate upper bound of number of rows.
@return Number of rows.
*/
ha_rows ha_partition::estimate_rows_upper_bound()
{
ha_rows rows, tot_rows= 0;
handler **file= m_file;
DBUG_ENTER("ha_partition::estimate_rows_upper_bound");
do
{
if (bitmap_is_set(&(m_part_info->read_partitions), (uint)(file - m_file)))
{
rows= (*file)->estimate_rows_upper_bound();
if (rows == HA_POS_ERROR)
DBUG_RETURN(HA_POS_ERROR);
tot_rows+= rows;
}
} while (*(++file));
DBUG_RETURN(tot_rows);
}
/*
Get time to read
SYNOPSIS
read_time()
index Index number used
ranges Number of ranges
rows Number of rows
RETURN VALUE
time for read
DESCRIPTION
This will be optimised later to include whether or not the index can
be used with partitioning. To achieve we need to add another parameter
that specifies how many of the index fields that are bound in the ranges.
Possibly added as a new call to handlers.
*/
double ha_partition::read_time(uint index, uint ranges, ha_rows rows)
{
DBUG_ENTER("ha_partition::read_time");
DBUG_RETURN(m_file[0]->read_time(index, ranges, rows));
}
/**
Number of rows in table. see handler.h
@return Number of records in the table (after pruning!)
*/
ha_rows ha_partition::records()
{
ha_rows rows, tot_rows= 0;
uint i;
DBUG_ENTER("ha_partition::records");
for (i= bitmap_get_first_set(&m_part_info->read_partitions);
i < m_tot_parts;
i= bitmap_get_next_set(&m_part_info->read_partitions, i))
{
rows= m_file[i]->records();
if (rows == HA_POS_ERROR)
DBUG_RETURN(HA_POS_ERROR);
tot_rows+= rows;
}
DBUG_RETURN(tot_rows);
}
/*
Is it ok to switch to a new engine for this table
SYNOPSIS
can_switch_engine()
RETURN VALUE
TRUE Ok
FALSE Not ok
DESCRIPTION
Used to ensure that tables with foreign key constraints are not moved
to engines without foreign key support.
*/
bool ha_partition::can_switch_engines()
{
handler **file;
DBUG_ENTER("ha_partition::can_switch_engines");
file= m_file;
do
{
if (!(*file)->can_switch_engines())
DBUG_RETURN(FALSE);
} while (*(++file));
DBUG_RETURN(TRUE);
}
/*
Is table cache supported
SYNOPSIS
table_cache_type()
*/
uint8 ha_partition::table_cache_type()
{
DBUG_ENTER("ha_partition::table_cache_type");
DBUG_RETURN(m_file[0]->table_cache_type());
}
/**
Calculate hash value for KEY partitioning using an array of fields.
@param field_array An array of the fields in KEY partitioning
@return hash_value calculated
@note Uses the hash function on the character set of the field.
Integer and floating point fields use the binary character set by default.
*/
uint32 ha_partition::calculate_key_hash_value(Field **field_array)
{
ulong nr1= 1;
ulong nr2= 4;
bool use_51_hash;
use_51_hash= MY_TEST((*field_array)->table->part_info->key_algorithm ==
partition_info::KEY_ALGORITHM_51);
do
{
Field *field= *field_array;
if (use_51_hash)
{
switch (field->real_type()) {
case MYSQL_TYPE_TINY:
case MYSQL_TYPE_SHORT:
case MYSQL_TYPE_LONG:
case MYSQL_TYPE_FLOAT:
case MYSQL_TYPE_DOUBLE:
case MYSQL_TYPE_NEWDECIMAL:
case MYSQL_TYPE_TIMESTAMP:
case MYSQL_TYPE_LONGLONG:
case MYSQL_TYPE_INT24:
case MYSQL_TYPE_TIME:
case MYSQL_TYPE_DATETIME:
case MYSQL_TYPE_YEAR:
case MYSQL_TYPE_NEWDATE:
{
if (field->is_null())
{
nr1^= (nr1 << 1) | 1;
continue;
}
/* Force this to my_hash_sort_bin, which was used in 5.1! */
uint len= field->pack_length();
my_charset_bin.coll->hash_sort(&my_charset_bin, field->ptr, len,
&nr1, &nr2);
/* Done with this field, continue with next one. */
continue;
}
case MYSQL_TYPE_STRING:
case MYSQL_TYPE_VARCHAR:
case MYSQL_TYPE_BIT:
/* Not affected, same in 5.1 and 5.5 */
break;
/*
ENUM/SET uses my_hash_sort_simple in 5.1 (i.e. my_charset_latin1)
and my_hash_sort_bin in 5.5!
*/
case MYSQL_TYPE_ENUM:
case MYSQL_TYPE_SET:
{
if (field->is_null())
{
nr1^= (nr1 << 1) | 1;
continue;
}
/* Force this to my_hash_sort_bin, which was used in 5.1! */
uint len= field->pack_length();
my_charset_latin1.coll->hash_sort(&my_charset_latin1, field->ptr,
len, &nr1, &nr2);
continue;
}
/* New types in mysql-5.6. */
case MYSQL_TYPE_DATETIME2:
case MYSQL_TYPE_TIME2:
case MYSQL_TYPE_TIMESTAMP2:
/* Not affected, 5.6+ only! */
break;
/* These types should not be allowed for partitioning! */
case MYSQL_TYPE_NULL:
case MYSQL_TYPE_DECIMAL:
case MYSQL_TYPE_DATE:
case MYSQL_TYPE_TINY_BLOB:
case MYSQL_TYPE_MEDIUM_BLOB:
case MYSQL_TYPE_LONG_BLOB:
case MYSQL_TYPE_BLOB:
case MYSQL_TYPE_VAR_STRING:
case MYSQL_TYPE_GEOMETRY:
/* fall through */
default:
DBUG_ASSERT(0); // New type?
/* Fall through for default hashing (5.5). */
}
/* fall through, use collation based hashing. */
}
field->hash(&nr1, &nr2);
} while (*(++field_array));
return (uint32) nr1;
}
/****************************************************************************
MODULE print messages
****************************************************************************/
const char *ha_partition::index_type(uint inx)
{
uint first_used_partition;
DBUG_ENTER("ha_partition::index_type");
first_used_partition= bitmap_get_first_set(&(m_part_info->read_partitions));
if (first_used_partition == MY_BIT_NONE)
{
DBUG_ASSERT(0); // How can this happen?
DBUG_RETURN(handler::index_type(inx));
}
DBUG_RETURN(m_file[first_used_partition]->index_type(inx));
}
enum row_type ha_partition::get_row_type() const
{
uint i;
enum row_type type;
DBUG_ENTER("ha_partition::get_row_type");
i= bitmap_get_first_set(&m_part_info->read_partitions);
DBUG_ASSERT(i < m_tot_parts);
if (i >= m_tot_parts)
DBUG_RETURN(ROW_TYPE_NOT_USED);
type= m_file[i]->get_row_type();
DBUG_PRINT("info", ("partition %u, row_type: %d", i, type));
for (i= bitmap_get_next_set(&m_part_info->lock_partitions, i);
i < m_tot_parts;
i= bitmap_get_next_set(&m_part_info->lock_partitions, i))
{
enum row_type part_type= m_file[i]->get_row_type();
DBUG_PRINT("info", ("partition %u, row_type: %d", i, type));
if (part_type != type)
DBUG_RETURN(ROW_TYPE_NOT_USED);
}
DBUG_RETURN(type);
}
void ha_partition::append_row_to_str(String &str)
{
const uchar *rec;
bool is_rec0= !m_err_rec || m_err_rec == table->record[0];
if (is_rec0)
rec= table->record[0];
else
rec= m_err_rec;
// If PK, use full PK instead of full part field array!
if (table->s->primary_key != MAX_KEY)
{
KEY *key= table->key_info + table->s->primary_key;
KEY_PART_INFO *key_part= key->key_part;
KEY_PART_INFO *key_part_end= key_part + key->user_defined_key_parts;
if (!is_rec0)
set_key_field_ptr(key, rec, table->record[0]);
for (; key_part != key_part_end; key_part++)
{
Field *field= key_part->field;
str.append(" ");
str.append(&field->field_name);
str.append(":");
field_unpack(&str, field, rec, 0, false);
}
if (!is_rec0)
set_key_field_ptr(key, table->record[0], rec);
}
else
{
Field **field_ptr;
if (!is_rec0)
table->move_fields(m_part_info->full_part_field_array, rec,
table->record[0]);
/* No primary key, use full partition field array. */
for (field_ptr= m_part_info->full_part_field_array;
*field_ptr;
field_ptr++)
{
Field *field= *field_ptr;
str.append(" ");
str.append(&field->field_name);
str.append(":");
field_unpack(&str, field, rec, 0, false);
}
if (!is_rec0)
table->move_fields(m_part_info->full_part_field_array, table->record[0],
rec);
}
}
void ha_partition::print_error(int error, myf errflag)
{
THD *thd= ha_thd();
DBUG_ENTER("ha_partition::print_error");
/* Should probably look for my own errors first */
DBUG_PRINT("enter", ("error: %d", error));
if ((error == HA_ERR_NO_PARTITION_FOUND) &&
! (thd->lex->alter_info.flags & Alter_info::ALTER_TRUNCATE_PARTITION))
{
m_part_info->print_no_partition_found(table, errflag);
DBUG_VOID_RETURN;
}
else if (error == HA_ERR_ROW_IN_WRONG_PARTITION)
{
/* Should only happen on DELETE or UPDATE! */
DBUG_ASSERT(thd_sql_command(thd) == SQLCOM_DELETE ||
thd_sql_command(thd) == SQLCOM_DELETE_MULTI ||
thd_sql_command(thd) == SQLCOM_UPDATE ||
thd_sql_command(thd) == SQLCOM_UPDATE_MULTI);
DBUG_ASSERT(m_err_rec);
if (m_err_rec)
{
uint max_length;
char buf[MAX_KEY_LENGTH];
String str(buf,sizeof(buf),system_charset_info);
uint32 part_id;
str.length(0);
str.append("(");
str.append_ulonglong(m_last_part);
str.append(" != ");
if (get_part_for_delete(m_err_rec, m_rec0, m_part_info, &part_id))
str.append("?");
else
str.append_ulonglong(part_id);
str.append(")");
append_row_to_str(str);
/* Log this error, so the DBA can notice it and fix it! */
sql_print_error("Table '%-192s' corrupted: row in wrong partition: %s\n"
"Please REPAIR the table!",
table->s->table_name.str,
str.c_ptr_safe());
max_length= (MYSQL_ERRMSG_SIZE -
(uint) strlen(ER_THD(thd, ER_ROW_IN_WRONG_PARTITION)));
if (str.length() >= max_length)
{
str.length(max_length-4);
str.append(STRING_WITH_LEN("..."));
}
my_error(ER_ROW_IN_WRONG_PARTITION, MYF(0), str.c_ptr_safe());
m_err_rec= NULL;
DBUG_VOID_RETURN;
}
/* fall through to generic error handling. */
}
/* In case m_file has not been initialized, like in bug#42438 */
if (m_file)
{
if (m_last_part >= m_tot_parts)
{
DBUG_ASSERT(0);
m_last_part= 0;
}
m_file[m_last_part]->print_error(error, errflag);
}
else
handler::print_error(error, errflag);
DBUG_VOID_RETURN;
}
bool ha_partition::get_error_message(int error, String *buf)
{
DBUG_ENTER("ha_partition::get_error_message");
/* Should probably look for my own errors first */
/* In case m_file has not been initialized, like in bug#42438 */
if (m_file)
DBUG_RETURN(m_file[m_last_part]->get_error_message(error, buf));
DBUG_RETURN(handler::get_error_message(error, buf));
}
/****************************************************************************
MODULE in-place ALTER
****************************************************************************/
/**
Get table flags.
*/
handler::Table_flags ha_partition::table_flags() const
{
uint first_used_partition= 0;
DBUG_ENTER("ha_partition::table_flags");
if (m_handler_status < handler_initialized ||
m_handler_status >= handler_closed)
DBUG_RETURN(PARTITION_ENABLED_TABLE_FLAGS);
if (get_lock_type() != F_UNLCK)
{
/*
The flags are cached after external_lock, and may depend on isolation
level. So we should use a locked partition to get the correct flags.
*/
first_used_partition= bitmap_get_first_set(&m_part_info->lock_partitions);
if (first_used_partition == MY_BIT_NONE)
first_used_partition= 0;
}
DBUG_RETURN((m_file[first_used_partition]->ha_table_flags() &
~(PARTITION_DISABLED_TABLE_FLAGS)) |
(PARTITION_ENABLED_TABLE_FLAGS));
}
/**
alter_table_flags must be on handler/table level, not on hton level
due to the ha_partition hton does not know what the underlying hton is.
*/
uint ha_partition::alter_table_flags(uint flags)
{
uint flags_to_return;
DBUG_ENTER("ha_partition::alter_table_flags");
flags_to_return= ht->alter_table_flags(flags);
flags_to_return|= m_file[0]->alter_table_flags(flags);
DBUG_RETURN(flags_to_return);
}
/**
check if copy of data is needed in alter table.
*/
bool ha_partition::check_if_incompatible_data(HA_CREATE_INFO *create_info,
uint table_changes)
{
handler **file;
bool ret= COMPATIBLE_DATA_YES;
/*
The check for any partitioning related changes have already been done
in mysql_alter_table (by fix_partition_func), so it is only up to
the underlying handlers.
*/
for (file= m_file; *file; file++)
if ((ret= (*file)->check_if_incompatible_data(create_info,
table_changes)) !=
COMPATIBLE_DATA_YES)
break;
return ret;
}
/**
Support of in-place alter table.
*/
/**
Helper class for in-place alter, see handler.h
*/
class ha_partition_inplace_ctx : public inplace_alter_handler_ctx
{
public:
inplace_alter_handler_ctx **handler_ctx_array;
private:
uint m_tot_parts;
public:
ha_partition_inplace_ctx(THD *thd, uint tot_parts)
: inplace_alter_handler_ctx(),
handler_ctx_array(NULL),
m_tot_parts(tot_parts)
{}
~ha_partition_inplace_ctx()
{
if (handler_ctx_array)
{
for (uint index= 0; index < m_tot_parts; index++)
delete handler_ctx_array[index];
}
}
};
enum_alter_inplace_result
ha_partition::check_if_supported_inplace_alter(TABLE *altered_table,
Alter_inplace_info *ha_alter_info)
{
uint index= 0;
enum_alter_inplace_result result= HA_ALTER_INPLACE_NO_LOCK;
ha_partition_inplace_ctx *part_inplace_ctx;
bool first_is_set= false;
THD *thd= ha_thd();
DBUG_ENTER("ha_partition::check_if_supported_inplace_alter");
/*
Support inplace change of KEY () -> KEY ALGORITHM = N ().
Any other change would set partition_changed in
prep_alter_part_table() in mysql_alter_table().
*/
if (ha_alter_info->alter_info->flags == Alter_info::ALTER_PARTITION)
DBUG_RETURN(HA_ALTER_INPLACE_NO_LOCK);
part_inplace_ctx=
new (thd->mem_root) ha_partition_inplace_ctx(thd, m_tot_parts);
if (!part_inplace_ctx)
DBUG_RETURN(HA_ALTER_ERROR);
part_inplace_ctx->handler_ctx_array= (inplace_alter_handler_ctx **)
thd->alloc(sizeof(inplace_alter_handler_ctx *) * (m_tot_parts + 1));
if (!part_inplace_ctx->handler_ctx_array)
DBUG_RETURN(HA_ALTER_ERROR);
/* Set all to NULL, including the terminating one. */
for (index= 0; index <= m_tot_parts; index++)
part_inplace_ctx->handler_ctx_array[index]= NULL;
ha_alter_info->handler_flags |= Alter_inplace_info::ALTER_PARTITIONED;
for (index= 0; index < m_tot_parts; index++)
{
enum_alter_inplace_result p_result=
m_file[index]->check_if_supported_inplace_alter(altered_table,
ha_alter_info);
part_inplace_ctx->handler_ctx_array[index]= ha_alter_info->handler_ctx;
if (index == 0)
{
first_is_set= (ha_alter_info->handler_ctx != NULL);
}
else if (first_is_set != (ha_alter_info->handler_ctx != NULL))
{
/* Either none or all partitions must set handler_ctx! */
DBUG_ASSERT(0);
DBUG_RETURN(HA_ALTER_ERROR);
}
if (p_result < result)
result= p_result;
if (result == HA_ALTER_ERROR)
break;
}
ha_alter_info->handler_ctx= part_inplace_ctx;
/*
To indicate for future inplace calls that there are several
partitions/handlers that need to be committed together,
we set group_commit_ctx to the NULL terminated array of
the partitions handlers.
*/
ha_alter_info->group_commit_ctx= part_inplace_ctx->handler_ctx_array;
DBUG_RETURN(result);
}
bool ha_partition::prepare_inplace_alter_table(TABLE *altered_table,
Alter_inplace_info *ha_alter_info)
{
uint index= 0;
bool error= false;
ha_partition_inplace_ctx *part_inplace_ctx;
DBUG_ENTER("ha_partition::prepare_inplace_alter_table");
/*
Changing to similar partitioning, only update metadata.
Non allowed changes would be catched in prep_alter_part_table().
*/
if (ha_alter_info->alter_info->flags == Alter_info::ALTER_PARTITION)
DBUG_RETURN(false);
part_inplace_ctx=
static_cast<class ha_partition_inplace_ctx*>(ha_alter_info->handler_ctx);
for (index= 0; index < m_tot_parts && !error; index++)
{
ha_alter_info->handler_ctx= part_inplace_ctx->handler_ctx_array[index];
if (m_file[index]->ha_prepare_inplace_alter_table(altered_table,
ha_alter_info))
error= true;
part_inplace_ctx->handler_ctx_array[index]= ha_alter_info->handler_ctx;
}
ha_alter_info->handler_ctx= part_inplace_ctx;
DBUG_RETURN(error);
}
bool ha_partition::inplace_alter_table(TABLE *altered_table,
Alter_inplace_info *ha_alter_info)
{
uint index= 0;
bool error= false;
ha_partition_inplace_ctx *part_inplace_ctx;
DBUG_ENTER("ha_partition::inplace_alter_table");
/*
Changing to similar partitioning, only update metadata.
Non allowed changes would be catched in prep_alter_part_table().
*/
if (ha_alter_info->alter_info->flags == Alter_info::ALTER_PARTITION)
DBUG_RETURN(false);
part_inplace_ctx=
static_cast<class ha_partition_inplace_ctx*>(ha_alter_info->handler_ctx);
for (index= 0; index < m_tot_parts && !error; index++)
{
ha_alter_info->handler_ctx= part_inplace_ctx->handler_ctx_array[index];
if (m_file[index]->ha_inplace_alter_table(altered_table,
ha_alter_info))
error= true;
part_inplace_ctx->handler_ctx_array[index]= ha_alter_info->handler_ctx;
}
ha_alter_info->handler_ctx= part_inplace_ctx;
DBUG_RETURN(error);
}
/*
Note that this function will try rollback failed ADD INDEX by
executing DROP INDEX for the indexes that were committed (if any)
before the error occurred. This means that the underlying storage
engine must be able to drop index in-place with X-lock held.
(As X-lock will be held here if new indexes are to be committed)
*/
bool ha_partition::commit_inplace_alter_table(TABLE *altered_table,
Alter_inplace_info *ha_alter_info,
bool commit)
{
ha_partition_inplace_ctx *part_inplace_ctx;
bool error= false;
DBUG_ENTER("ha_partition::commit_inplace_alter_table");
/*
Changing to similar partitioning, only update metadata.
Non allowed changes would be catched in prep_alter_part_table().
*/
if (ha_alter_info->alter_info->flags == Alter_info::ALTER_PARTITION)
DBUG_RETURN(false);
part_inplace_ctx=
static_cast<class ha_partition_inplace_ctx*>(ha_alter_info->handler_ctx);
if (commit)
{
DBUG_ASSERT(ha_alter_info->group_commit_ctx ==
part_inplace_ctx->handler_ctx_array);
ha_alter_info->handler_ctx= part_inplace_ctx->handler_ctx_array[0];
error= m_file[0]->ha_commit_inplace_alter_table(altered_table,
ha_alter_info, commit);
if (error)
goto end;
if (ha_alter_info->group_commit_ctx)
{
/*
If ha_alter_info->group_commit_ctx is not set to NULL,
then the engine did only commit the first partition!
The engine is probably new, since both innodb and the default
implementation of handler::commit_inplace_alter_table sets it to NULL
and simply return false, since it allows metadata changes only.
Loop over all other partitions as to follow the protocol!
*/
uint i;
DBUG_ASSERT(0);
for (i= 1; i < m_tot_parts; i++)
{
ha_alter_info->handler_ctx= part_inplace_ctx->handler_ctx_array[i];
error|= m_file[i]->ha_commit_inplace_alter_table(altered_table,
ha_alter_info,
true);
}
}
}
else
{
uint i;
for (i= 0; i < m_tot_parts; i++)
{
/* Rollback, commit == false, is done for each partition! */
ha_alter_info->handler_ctx= part_inplace_ctx->handler_ctx_array[i];
if (m_file[i]->ha_commit_inplace_alter_table(altered_table,
ha_alter_info, false))
error= true;
}
}
end:
ha_alter_info->handler_ctx= part_inplace_ctx;
DBUG_RETURN(error);
}
void ha_partition::notify_table_changed()
{
handler **file;
DBUG_ENTER("ha_partition::notify_table_changed");
for (file= m_file; *file; file++)
(*file)->ha_notify_table_changed();
DBUG_VOID_RETURN;
}
uint ha_partition::min_of_the_max_uint(
uint (handler::*operator_func)(void) const) const
{
handler **file;
uint min_of_the_max= ((*m_file)->*operator_func)();
for (file= m_file+1; *file; file++)
{
uint tmp= ((*file)->*operator_func)();
set_if_smaller(min_of_the_max, tmp);
}
return min_of_the_max;
}
uint ha_partition::max_supported_key_parts() const
{
return min_of_the_max_uint(&handler::max_supported_key_parts);
}
uint ha_partition::max_supported_key_length() const
{
return min_of_the_max_uint(&handler::max_supported_key_length);
}
uint ha_partition::max_supported_key_part_length() const
{
return min_of_the_max_uint(&handler::max_supported_key_part_length);
}
uint ha_partition::max_supported_record_length() const
{
return min_of_the_max_uint(&handler::max_supported_record_length);
}
uint ha_partition::max_supported_keys() const
{
return min_of_the_max_uint(&handler::max_supported_keys);
}
uint ha_partition::min_record_length(uint options) const
{
handler **file;
uint max= (*m_file)->min_record_length(options);
for (file= m_file, file++; *file; file++)
if (max < (*file)->min_record_length(options))
max= (*file)->min_record_length(options);
return max;
}
/****************************************************************************
MODULE compare records
****************************************************************************/
/*
Compare two positions
SYNOPSIS
cmp_ref()
ref1 First position
ref2 Second position
RETURN VALUE
<0 ref1 < ref2
0 Equal
>0 ref1 > ref2
DESCRIPTION
We get two references and need to check if those records are the same.
If they belong to different partitions we decide that they are not
the same record. Otherwise we use the particular handler to decide if
they are the same. Sort in partition id order if not equal.
MariaDB note:
Please don't merge the code from MySQL that does this:
We get two references and need to check if those records are the same.
If they belong to different partitions we decide that they are not
the same record. Otherwise we use the particular handler to decide if
they are the same. Sort in partition id order if not equal.
It is incorrect, MariaDB has an alternative fix.
*/
int ha_partition::cmp_ref(const uchar *ref1, const uchar *ref2)
{
int cmp;
my_ptrdiff_t diff1, diff2;
DBUG_ENTER("ha_partition::cmp_ref");
cmp = m_file[0]->cmp_ref((ref1 + PARTITION_BYTES_IN_POS),
(ref2 + PARTITION_BYTES_IN_POS));
if (cmp)
DBUG_RETURN(cmp);
if ((ref1[0] == ref2[0]) && (ref1[1] == ref2[1]))
{
/* This means that the references are same and are in same partition.*/
DBUG_RETURN(0);
}
/*
In Innodb we compare with either primary key value or global DB_ROW_ID so
it is not possible that the two references are equal and are in different
partitions, but in myisam it is possible since we are comparing offsets.
Remove this assert if DB_ROW_ID is changed to be per partition.
*/
DBUG_ASSERT(!m_innodb);
diff1= ref2[1] - ref1[1];
diff2= ref2[0] - ref1[0];
if (diff1 > 0)
{
DBUG_RETURN(-1);
}
if (diff1 < 0)
{
DBUG_RETURN(+1);
}
if (diff2 > 0)
{
DBUG_RETURN(-1);
}
DBUG_RETURN(+1);
}
/****************************************************************************
MODULE auto increment
****************************************************************************/
int ha_partition::reset_auto_increment(ulonglong value)
{
handler **file= m_file;
int res;
DBUG_ENTER("ha_partition::reset_auto_increment");
lock_auto_increment();
part_share->auto_inc_initialized= false;
part_share->next_auto_inc_val= 0;
do
{
if ((res= (*file)->ha_reset_auto_increment(value)) != 0)
break;
} while (*(++file));
unlock_auto_increment();
DBUG_RETURN(res);
}
/**
This method is called by update_auto_increment which in turn is called
by the individual handlers as part of write_row. We use the
part_share->next_auto_inc_val, or search all
partitions for the highest auto_increment_value if not initialized or
if auto_increment field is a secondary part of a key, we must search
every partition when holding a mutex to be sure of correctness.
*/
void ha_partition::get_auto_increment(ulonglong offset, ulonglong increment,
ulonglong nb_desired_values,
ulonglong *first_value,
ulonglong *nb_reserved_values)
{
DBUG_ENTER("ha_partition::get_auto_increment");
DBUG_PRINT("info", ("offset: %lu inc: %lu desired_values: %lu "
"first_value: %lu", (ulong) offset, (ulong) increment,
(ulong) nb_desired_values, (ulong) *first_value));
DBUG_ASSERT(increment && nb_desired_values);
*first_value= 0;
if (table->s->next_number_keypart)
{
/*
next_number_keypart is != 0 if the auto_increment column is a secondary
column in the index (it is allowed in MyISAM)
*/
DBUG_PRINT("info", ("next_number_keypart != 0"));
ulonglong nb_reserved_values_part;
ulonglong first_value_part, max_first_value;
handler **file= m_file;
first_value_part= max_first_value= *first_value;
/* Must find highest value among all partitions. */
do
{
/* Only nb_desired_values = 1 makes sense */
(*file)->get_auto_increment(offset, increment, 1,
&first_value_part, &nb_reserved_values_part);
if (first_value_part == ULONGLONG_MAX) // error in one partition
{
*first_value= first_value_part;
/* log that the error was between table/partition handler */
sql_print_error("Partition failed to reserve auto_increment value");
DBUG_VOID_RETURN;
}
DBUG_PRINT("info", ("first_value_part: %lu", (ulong) first_value_part));
set_if_bigger(max_first_value, first_value_part);
} while (*(++file));
*first_value= max_first_value;
*nb_reserved_values= 1;
}
else
{
THD *thd= ha_thd();
/*
This is initialized in the beginning of the first write_row call.
*/
DBUG_ASSERT(part_share->auto_inc_initialized);
/*
Get a lock for handling the auto_increment in part_share
for avoiding two concurrent statements getting the same number.
*/
lock_auto_increment();
/*
In a multi-row insert statement like INSERT SELECT and LOAD DATA
where the number of candidate rows to insert is not known in advance
we must hold a lock/mutex for the whole statement if we have statement
based replication. Because the statement-based binary log contains
only the first generated value used by the statement, and slaves assumes
all other generated values used by this statement were consecutive to
this first one, we must exclusively lock the generator until the statement
is done.
*/
if (!auto_increment_safe_stmt_log_lock &&
thd->lex->sql_command != SQLCOM_INSERT &&
mysql_bin_log.is_open() &&
!thd->is_current_stmt_binlog_format_row() &&
(thd->variables.option_bits & OPTION_BIN_LOG))
{
DBUG_PRINT("info", ("locking auto_increment_safe_stmt_log_lock"));
auto_increment_safe_stmt_log_lock= TRUE;
}
/* this gets corrected (for offset/increment) in update_auto_increment */
*first_value= part_share->next_auto_inc_val;
part_share->next_auto_inc_val+= nb_desired_values * increment;
unlock_auto_increment();
DBUG_PRINT("info", ("*first_value: %lu", (ulong) *first_value));
*nb_reserved_values= nb_desired_values;
}
DBUG_VOID_RETURN;
}
void ha_partition::release_auto_increment()
{
DBUG_ENTER("ha_partition::release_auto_increment");
if (table->s->next_number_keypart)
{
uint i;
for (i= bitmap_get_first_set(&m_part_info->lock_partitions);
i < m_tot_parts;
i= bitmap_get_next_set(&m_part_info->lock_partitions, i))
{
m_file[i]->ha_release_auto_increment();
}
}
else if (next_insert_id)
{
ulonglong next_auto_inc_val;
lock_auto_increment();
next_auto_inc_val= part_share->next_auto_inc_val;
/*
If the current auto_increment values is lower than the reserved
value, and the reserved value was reserved by this thread,
we can lower the reserved value.
*/
if (next_insert_id < next_auto_inc_val &&
auto_inc_interval_for_cur_row.maximum() >= next_auto_inc_val)
{
THD *thd= ha_thd();
/*
Check that we do not lower the value because of a failed insert
with SET INSERT_ID, i.e. forced/non generated values.
*/
if (thd->auto_inc_intervals_forced.maximum() < next_insert_id)
part_share->next_auto_inc_val= next_insert_id;
}
DBUG_PRINT("info", ("part_share->next_auto_inc_val: %lu",
(ulong) part_share->next_auto_inc_val));
/* Unlock the multi row statement lock taken in get_auto_increment */
if (auto_increment_safe_stmt_log_lock)
{
auto_increment_safe_stmt_log_lock= FALSE;
DBUG_PRINT("info", ("unlocking auto_increment_safe_stmt_log_lock"));
}
unlock_auto_increment();
}
DBUG_VOID_RETURN;
}
/****************************************************************************
MODULE initialize handler for HANDLER call
****************************************************************************/
void ha_partition::init_table_handle_for_HANDLER()
{
return;
}
/**
Return the checksum of the table (all partitions)
*/
uint ha_partition::checksum() const
{
ha_checksum sum= 0;
DBUG_ENTER("ha_partition::checksum");
if ((table_flags() & (HA_HAS_OLD_CHECKSUM | HA_HAS_NEW_CHECKSUM)))
{
handler **file= m_file;
do
{
sum+= (*file)->checksum();
} while (*(++file));
}
DBUG_RETURN(sum);
}
/****************************************************************************
MODULE enable/disable indexes
****************************************************************************/
/*
Disable indexes for a while
SYNOPSIS
disable_indexes()
mode Mode
RETURN VALUES
0 Success
!= 0 Error
*/
int ha_partition::disable_indexes(uint mode)
{
handler **file;
int error= 0;
DBUG_ASSERT(bitmap_is_set_all(&(m_part_info->lock_partitions)));
for (file= m_file; *file; file++)
{
if ((error= (*file)->ha_disable_indexes(mode)))
break;
}
return error;
}
/*
Enable indexes again
SYNOPSIS
enable_indexes()
mode Mode
RETURN VALUES
0 Success
!= 0 Error
*/
int ha_partition::enable_indexes(uint mode)
{
handler **file;
int error= 0;
DBUG_ASSERT(bitmap_is_set_all(&(m_part_info->lock_partitions)));
for (file= m_file; *file; file++)
{
if ((error= (*file)->ha_enable_indexes(mode)))
break;
}
return error;
}
/*
Check if indexes are disabled
SYNOPSIS
indexes_are_disabled()
RETURN VALUES
0 Indexes are enabled
!= 0 Indexes are disabled
*/
int ha_partition::indexes_are_disabled(void)
{
handler **file;
int error= 0;
DBUG_ASSERT(bitmap_is_set_all(&(m_part_info->lock_partitions)));
for (file= m_file; *file; file++)
{
if ((error= (*file)->indexes_are_disabled()))
break;
}
return error;
}
/**
Check/fix misplaced rows.
@param read_part_id Partition to check/fix.
@param repair If true, move misplaced rows to correct partition.
@return Operation status.
@retval 0 Success
@retval != 0 Error
*/
int ha_partition::check_misplaced_rows(uint read_part_id, bool do_repair)
{
int result= 0;
uint32 correct_part_id;
longlong func_value;
longlong num_misplaced_rows= 0;
DBUG_ENTER("ha_partition::check_misplaced_rows");
DBUG_ASSERT(m_file);
if (do_repair)
{
/* We must read the full row, if we need to move it! */
bitmap_set_all(table->read_set);
bitmap_set_all(table->write_set);
}
else
{
/* Only need to read the partitioning fields. */
bitmap_union(table->read_set, &m_part_info->full_part_field_set);
}
if ((result= m_file[read_part_id]->ha_rnd_init(1)))
DBUG_RETURN(result);
while (true)
{
if ((result= m_file[read_part_id]->ha_rnd_next(m_rec0)))
{
if (result == HA_ERR_RECORD_DELETED)
continue;
if (result != HA_ERR_END_OF_FILE)
break;
if (num_misplaced_rows > 0)
{
print_admin_msg(ha_thd(), MYSQL_ERRMSG_SIZE, "warning",
table_share->db.str, table->alias,
opt_op_name[REPAIR_PARTS],
"Moved %lld misplaced rows",
num_misplaced_rows);
}
/* End-of-file reached, all rows are now OK, reset result and break. */
result= 0;
break;
}
result= m_part_info->get_partition_id(m_part_info, &correct_part_id,
&func_value);
if (result)
break;
if (correct_part_id != read_part_id)
{
num_misplaced_rows++;
if (!do_repair)
{
/* Check. */
print_admin_msg(ha_thd(), MYSQL_ERRMSG_SIZE, "error",
table_share->db.str, table->alias,
opt_op_name[CHECK_PARTS],
"Found a misplaced row");
/* Break on first misplaced row! */
result= HA_ADMIN_NEEDS_UPGRADE;
break;
}
else
{
DBUG_PRINT("info", ("Moving row from partition %d to %d",
read_part_id, correct_part_id));
/*
Insert row into correct partition. Notice that there are no commit
for every N row, so the repair will be one large transaction!
*/
if ((result= m_file[correct_part_id]->ha_write_row(m_rec0)))
{
/*
We have failed to insert a row, it might have been a duplicate!
*/
char buf[MAX_KEY_LENGTH];
String str(buf,sizeof(buf),system_charset_info);
str.length(0);
if (result == HA_ERR_FOUND_DUPP_KEY)
{
str.append("Duplicate key found, "
"please update or delete the record:\n");
result= HA_ADMIN_CORRUPT;
}
m_err_rec= NULL;
append_row_to_str(str);
/*
If the engine supports transactions, the failure will be
rollbacked.
*/
if (!m_file[correct_part_id]->has_transactions())
{
/* Log this error, so the DBA can notice it and fix it! */
sql_print_error("Table '%-192s' failed to move/insert a row"
" from part %d into part %d:\n%s",
table->s->table_name.str,
read_part_id,
correct_part_id,
str.c_ptr_safe());
}
print_admin_msg(ha_thd(), MYSQL_ERRMSG_SIZE, "error",
table_share->db.str, table->alias,
opt_op_name[REPAIR_PARTS],
"Failed to move/insert a row"
" from part %d into part %d:\n%s",
read_part_id,
correct_part_id,
str.c_ptr_safe());
break;
}
/* Delete row from wrong partition. */
if ((result= m_file[read_part_id]->ha_delete_row(m_rec0)))
{
if (m_file[correct_part_id]->has_transactions())
break;
/*
We have introduced a duplicate, since we failed to remove it
from the wrong partition.
*/
char buf[MAX_KEY_LENGTH];
String str(buf,sizeof(buf),system_charset_info);
str.length(0);
m_err_rec= NULL;
append_row_to_str(str);
/* Log this error, so the DBA can notice it and fix it! */
sql_print_error("Table '%-192s': Delete from part %d failed with"
" error %d. But it was already inserted into"
" part %d, when moving the misplaced row!"
"\nPlease manually fix the duplicate row:\n%s",
table->s->table_name.str,
read_part_id,
result,
correct_part_id,
str.c_ptr_safe());
break;
}
}
}
}
int tmp_result= m_file[read_part_id]->ha_rnd_end();
DBUG_RETURN(result ? result : tmp_result);
}
#define KEY_PARTITIONING_CHANGED_STR \
"KEY () partitioning changed, please run:\n" \
"ALTER TABLE %s.%s ALGORITHM = INPLACE %s"
int ha_partition::check_for_upgrade(HA_CHECK_OPT *check_opt)
{
int error= HA_ADMIN_NEEDS_CHECK;
DBUG_ENTER("ha_partition::check_for_upgrade");
/*
This is called even without FOR UPGRADE,
if the .frm version is lower than the current version.
In that case return that it needs checking!
*/
if (!(check_opt->sql_flags & TT_FOR_UPGRADE))
DBUG_RETURN(error);
/*
Partitions will be checked for during their ha_check!
Check if KEY (sub)partitioning was used and any field's hash calculation
differs from 5.1, see bug#14521864.
*/
if (table->s->mysql_version < 50503 && // 5.1 table (<5.5.3)
((m_part_info->part_type == HASH_PARTITION && // KEY partitioned
m_part_info->list_of_part_fields) ||
(m_is_sub_partitioned && // KEY subpartitioned
m_part_info->list_of_subpart_fields)))
{
Field **field;
if (m_is_sub_partitioned)
{
field= m_part_info->subpart_field_array;
}
else
{
field= m_part_info->part_field_array;
}
for (; *field; field++)
{
switch ((*field)->real_type()) {
case MYSQL_TYPE_TINY:
case MYSQL_TYPE_SHORT:
case MYSQL_TYPE_LONG:
case MYSQL_TYPE_FLOAT:
case MYSQL_TYPE_DOUBLE:
case MYSQL_TYPE_NEWDECIMAL:
case MYSQL_TYPE_TIMESTAMP:
case MYSQL_TYPE_LONGLONG:
case MYSQL_TYPE_INT24:
case MYSQL_TYPE_TIME:
case MYSQL_TYPE_DATETIME:
case MYSQL_TYPE_YEAR:
case MYSQL_TYPE_NEWDATE:
case MYSQL_TYPE_ENUM:
case MYSQL_TYPE_SET:
{
THD *thd= ha_thd();
char *part_buf;
String db_name, table_name;
uint part_buf_len;
bool skip_generation= false;
partition_info::enum_key_algorithm old_algorithm;
old_algorithm= m_part_info->key_algorithm;
error= HA_ADMIN_FAILED;
append_identifier(ha_thd(), &db_name, table_share->db.str,
table_share->db.length);
append_identifier(ha_thd(), &table_name, table_share->table_name.str,
table_share->table_name.length);
if (m_part_info->key_algorithm != partition_info::KEY_ALGORITHM_NONE)
{
/*
Only possible when someone tampered with .frm files,
like during tests :)
*/
skip_generation= true;
}
m_part_info->key_algorithm= partition_info::KEY_ALGORITHM_51;
if (skip_generation ||
!(part_buf= generate_partition_syntax(thd, m_part_info,
&part_buf_len,
true,
NULL,
NULL)) ||
print_admin_msg(thd, SQL_ADMIN_MSG_TEXT_SIZE + 1, "error",
table_share->db.str,
table->alias,
opt_op_name[CHECK_PARTS],
KEY_PARTITIONING_CHANGED_STR,
db_name.c_ptr_safe(),
table_name.c_ptr_safe(),
part_buf))
{
/* Error creating admin message (too long string?). */
print_admin_msg(thd, MYSQL_ERRMSG_SIZE, "error",
table_share->db.str, table->alias,
opt_op_name[CHECK_PARTS],
KEY_PARTITIONING_CHANGED_STR,
db_name.c_ptr_safe(), table_name.c_ptr_safe(),
"<old partition clause>, but add ALGORITHM = 1"
" between 'KEY' and '(' to change the metadata"
" without the need of a full table rebuild.");
}
m_part_info->key_algorithm= old_algorithm;
DBUG_RETURN(error);
}
default:
/* Not affected! */
;
}
}
}
DBUG_RETURN(error);
}
struct st_mysql_storage_engine partition_storage_engine=
{ MYSQL_HANDLERTON_INTERFACE_VERSION };
maria_declare_plugin(partition)
{
MYSQL_STORAGE_ENGINE_PLUGIN,
&partition_storage_engine,
"partition",
"Mikael Ronstrom, MySQL AB",
"Partition Storage Engine Helper",
PLUGIN_LICENSE_GPL,
partition_initialize, /* Plugin Init */
NULL, /* Plugin Deinit */
0x0100, /* 1.0 */
NULL, /* status variables */
NULL, /* system variables */
"1.0", /* string version */
MariaDB_PLUGIN_MATURITY_STABLE /* maturity */
}
maria_declare_plugin_end;
#endif
|