add awful but working sampling method selection

6 files changed, 53 insertions, 5 deletions

diff --git a/sql/lex.h b/sql/lex.h
index 55589d888e6..b2ca572481d 100644
--- a/sql/lex.h
+++ b/sql/lex.h
@@ -571,6 +571,7 @@ SYMBOL symbols[] = {
   /** sql_function and condition_property_name for GET DIAGNOSTICS */
   { "ROW_NUMBER",	SYM(ROW_NUMBER_SYM)},
   { "RTREE",		SYM(RTREE_SYM)},
+    {"SAMPLE_METHOD", SYM(SAMPLE_METHOD)},
   { "SAVEPOINT",	SYM(SAVEPOINT_SYM)},
   { "SCHEDULE",		SYM(SCHEDULE_SYM)},
   { "SCHEMA",		SYM(DATABASE)},
diff --git a/sql/records.cc b/sql/records.cc
index 9eadf788585..896ecafe1c0 100644
--- a/sql/records.cc
+++ b/sql/records.cc
@@ -38,6 +38,7 @@
 static int rr_quick(READ_RECORD *info);
 int rr_sequential(READ_RECORD *info);
 int rr_sequential_sample(READ_RECORD *info);
+int rr_full_scan_sample(READ_RECORD *info);
 static int rr_from_tempfile(READ_RECORD *info);
 template<bool> static int rr_unpack_from_tempfile(READ_RECORD *info);
 template<bool,bool> static int rr_unpack_from_buffer(READ_RECORD *info);
@@ -325,10 +326,26 @@ bool init_read_record(READ_RECORD *info,THD *thd, TABLE *table,
   else if (table->tablesample)
   {
     double fract= table->tablesample->val_real() / 100.0;
+    info->sample_factor = fract;
     info->sample_counter= (ha_rows)(table->file->records() * fract + 0.5);
-    info->read_record_func= rr_sequential_sample;
-    if (table->file->ha_sample_init())
-      DBUG_RETURN(1);
+    if(table->sample_method_flag && table->sample_method_flag->val_real() > 1.0)
+    {
+      info->read_record_func= rr_full_scan_sample;
+    }
+    else
+    {
+      info->read_record_func= rr_sequential_sample;
+    }
+    if(table->sample_method_flag && table->sample_method_flag->val_real() > 1.0)
+    {
+      if (table->file->ha_rnd_init(TRUE))
+        DBUG_RETURN(1);
+    }
+    else
+    {
+      if (table->file->ha_sample_init())
+        DBUG_RETURN(1);
+    }
   }
   else
   {
@@ -548,6 +565,20 @@ int rr_sequential_sample(READ_RECORD *info)
   return tmp;
 }
 
+int rr_full_scan_sample(READ_RECORD *info)
+{
+  int tmp;
+  do
+  {
+    tmp= info->table->file->ha_rnd_next(info->record());
+    if(thd_rnd(info->thd) <= info->sample_factor)
+      break;
+  } while (!tmp);
+  if(tmp)
+    tmp= rr_handle_error(info, tmp);
+  return tmp;
+}
+
 
 static int rr_from_tempfile(READ_RECORD *info)
 {
diff --git a/sql/records.h b/sql/records.h
index 579e0e2bf3a..ed3b9a84b5a 100644
--- a/sql/records.h
+++ b/sql/records.h
@@ -62,6 +62,7 @@ struct READ_RECORD
   uint ref_length, reclength, rec_cache_size, error_offset;
 
   ha_rows sample_counter;
+  double sample_factor;
   /**
     Counting records when reading result from filesort().
     Used when filesort leaves the result in the filesort buffer.
diff --git a/sql/sql_yacc.yy b/sql/sql_yacc.yy
index 88df2bc3d7a..0399225ad5f 100644
--- a/sql/sql_yacc.yy
+++ b/sql/sql_yacc.yy
@@ -679,6 +679,7 @@ bool my_yyoverflow(short **a, YYSTYPE **b, size_t *yystacksize);
 %token  <kwd> SYSDATE
 %token  <kwd> TABLE_REF_PRIORITY
 %token  <kwd> TABLESAMPLE                   /* SQL-2016-R */
+%token  <kwd> SAMPLE_METHOD
 %token  <kwd> TABLE_SYM                     /* SQL-2003-R */
 %token  <kwd> TERMINATED
 %token  <kwd> THEN_SYM                      /* SQL-2003-R */
@@ -1499,7 +1500,7 @@ bool my_yyoverflow(short **a, YYSTYPE **b, size_t *yystacksize);
 %type <item_param> param_marker
 
 %type <item_num>
-        NUM_literal opt_table_sample
+        NUM_literal opt_table_sample opt_sample_method
 
 %type <item_basic_constant> text_literal
 
@@ -11710,10 +11711,18 @@ opt_table_sample:
         }
         ;
 
+opt_sample_method:
+        /* empty */ { $$=0; }
+        | SAMPLE_METHOD '(' NUM_literal ')'
+        {
+          $$=$3;
+        }
+        ;
+
 table_primary_ident:
           table_ident opt_use_partition opt_for_system_time_clause
           opt_table_alias_clause opt_key_definition
-          opt_table_sample
+          opt_table_sample opt_sample_method
           {
             if (!($$= Select->add_table_to_list(thd, $1, $4,
                                                 0,
@@ -11726,6 +11735,8 @@ table_primary_ident:
               $$->vers_conditions= Lex->vers_conditions;
             if ($6)
               $$->tablesample= $6;
+            if ($7)
+              $$->sample_method_flag= $7;
           }
         ;
 
diff --git a/sql/table.cc b/sql/table.cc
index 4504e37dd7f..795a9e0b754 100644
--- a/sql/table.cc
+++ b/sql/table.cc
@@ -5555,6 +5555,7 @@ void TABLE::init(THD *thd, TABLE_LIST *tl)
   no_cache= false;
   initialize_opt_range_structures();
   tablesample= NULL;
+  sample_method_flag= NULL;
 #ifdef HAVE_REPLICATION
   /* used in RBR Triggers */
   master_had_triggers= 0;
@@ -8546,6 +8547,7 @@ bool TABLE_LIST::process_table_sample(TABLE *tbl)
     tbl->keys_in_use_for_order_by.clear_all();
     tbl->covering_keys.clear_all();
     tbl->tablesample= tablesample;
+    tbl->sample_method_flag= sample_method_flag;
   }
   return false;
 }
diff --git a/sql/table.h b/sql/table.h
index 5c0e7f9b51a..266249fd9aa 100644
--- a/sql/table.h
+++ b/sql/table.h
@@ -1498,6 +1498,7 @@ public:
   bool master_had_triggers;
 #endif
   Item *tablesample;
+  Item *sample_method_flag; /* rand_sample if>=1.0, full_scan_sample another*/
   REGINFO reginfo;			/* field connections */
   MEM_ROOT mem_root;
   /**
@@ -2223,6 +2224,7 @@ struct TABLE_LIST
     for_insert_data= insert_data;
   }
   Item *tablesample;
+  Item *sample_method_flag;
 
   /*
     List of tables local to a subquery (used by SQL_I_List). Considers