Bug#46639: 1030 (HY000): Got error 124 from storage engine on

INSERT ... SELECT ...

Problem was that when bulk insert is used on an empty
table/partition, it disables the indexes for better
performance, but in this specific case it also tries
to read from that partition using an index, which is
not possible since it has been disabled.

Solution was to allow index reads on disabled indexes
if there are no records.

Also reverted the patch for bug#38005, since that was a workaround
in the partitioning engine instead of a fix in myisam.

mysql-test/r/partition.result:
  Bug#46639: 1030 (HY000): Got error 124 from storage engine on
  INSERT ... SELECT ...
  
  updated result file
mysql-test/t/partition.test:
  Bug#46639: 1030 (HY000): Got error 124 from storage engine on
  INSERT ... SELECT ...
  
  Added testcase
sql/ha_partition.cc:
  Bug#46639: 1030 (HY000): Got error 124 from storage engine on
  INSERT ... SELECT ...
  
  reverted the patch for bug#38005, since that was a workaround
  around this problem, not needed after fixing it in myisam.
storage/myisam/mi_search.c:
  Bug#46639: 1030 (HY000): Got error 124 from storage engine on
  INSERT ... SELECT ...
  
  Return HA_ERR_END_OF_FILE instead of HA_ERR_WRONG_INDEX
  when there are no rows.

1 files changed, 0 insertions, 33 deletions

diff --git a/sql/ha_partition.cc b/sql/ha_partition.cc
index b345648260c..ac8c46ec4e3 100644
--- a/sql/ha_partition.cc
+++ b/sql/ha_partition.cc
@@ -4414,17 +4414,6 @@ int ha_partition::handle_unordered_scan_next_partition(uchar * buf)
       break;
     case partition_index_first:
       DBUG_PRINT("info", ("index_first on partition %d", i));
-      /* MyISAM engine can fail if we call index_first() when indexes disabled */
-      /* that happens if the table is empty. */
-      /* Here we use file->stats.records instead of file->records() because */
-      /* file->records() is supposed to return an EXACT count, and it can be   */
-      /* possibly slow. We don't need an exact number, an approximate one- from*/
-      /* the last ::info() call - is sufficient. */
-      if (file->stats.records == 0)
-      {
-        error= HA_ERR_END_OF_FILE;
-        break;
-      }
       error= file->index_first(buf);
       break;
     case partition_index_first_unordered:
@@ -4512,32 +4501,10 @@ int ha_partition::handle_ordered_index_scan(uchar *buf, bool reverse_order)
                                   m_start_key.flag);
       break;
     case partition_index_first:
-      /* MyISAM engine can fail if we call index_first() when indexes disabled */
-      /* that happens if the table is empty. */
-      /* Here we use file->stats.records instead of file->records() because */
-      /* file->records() is supposed to return an EXACT count, and it can be   */
-      /* possibly slow. We don't need an exact number, an approximate one- from*/
-      /* the last ::info() call - is sufficient. */
-      if (file->stats.records == 0)
-      {
-        error= HA_ERR_END_OF_FILE;
-        break;
-      }
       error= file->index_first(rec_buf_ptr);
       reverse_order= FALSE;
       break;
     case partition_index_last:
-      /* MyISAM engine can fail if we call index_last() when indexes disabled */
-      /* that happens if the table is empty. */
-      /* Here we use file->stats.records instead of file->records() because */
-      /* file->records() is supposed to return an EXACT count, and it can be   */
-      /* possibly slow. We don't need an exact number, an approximate one- from*/
-      /* the last ::info() call - is sufficient. */
-      if (file->stats.records == 0)
-      {
-        error= HA_ERR_END_OF_FILE;
-        break;
-      }
       error= file->index_last(rec_buf_ptr);
       reverse_order= TRUE;
       break;