optimisation of independent ALL/ANY with aggregate function (WL#1115) (SCRUM)

mysql-test/r/subselect.result

@@ -1347,4 +1347,13 @@ explain select * from t3 where a >= all (select b from t2);
 id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
 1	PRIMARY	t3	ALL	NULL	NULL	NULL	NULL	3	Using where
 2	SUBQUERY	t2	system	NULL	NULL	NULL	NULL	0	const row not found
+insert into t2 values (2,2), (2,1), (3,3), (3,1);
+select * from t3 where a > all (select max(b) from t2 group by a);
+a
+6
+7
+explain select * from t3 where a > all (select max(b) from t2 group by a);
+id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
+1	PRIMARY	t3	ALL	NULL	NULL	NULL	NULL	3	Using where
+2	SUBQUERY	t2	ALL	NULL	NULL	NULL	NULL	4	Using temporary; Using filesort
 drop table if exists t2, t3;

mysql-test/t/subselect.test

@@ -877,6 +877,7 @@ insert into t1 values (1);
 insert into t2 values (1);
 select * from t1 where exists (select s1 from t2 having max(t2.s1)=t1.s1);
 drop table t1,t2;
+
 #
 # update subquery with wrong field (to force name resolving
 # in UPDATE name space)
@@ -897,4 +898,9 @@ create table t3 (a int);
 insert into t3 values (6),(7),(3);
 select * from t3 where a >= all (select b from t2);
 explain select * from t3 where a >= all (select b from t2);
+
+# optimized static ALL/ANY with grouping
+insert into t2 values (2,2), (2,1), (3,3), (3,1);
+select * from t3 where a > all (select max(b) from t2 group by a);
+explain select * from t3 where a > all (select max(b) from t2 group by a);
 drop table if exists t2, t3;

sql/item_subselect.cc

@@ -152,8 +152,8 @@ inline table_map Item_subselect::used_tables() const
 }
 
 Item_singlerow_subselect::Item_singlerow_subselect(THD *thd,
-						   st_select_lex *select_lex):
-  Item_subselect(), value(0)
+						   st_select_lex *select_lex)
+  :Item_subselect(), value(0)
 {
   DBUG_ENTER("Item_singlerow_subselect::Item_singlerow_subselect");
   init(thd, select_lex, new select_singlerow_subselect(this));
@@ -163,6 +163,19 @@ Item_singlerow_subselect::Item_singlerow_subselect(THD *thd,
   DBUG_VOID_RETURN;
 }
 
+Item_maxmin_subselect::Item_maxmin_subselect(THD *thd,
+					     st_select_lex *select_lex,
+					     bool max)
+  :Item_singlerow_subselect()
+{
+  DBUG_ENTER("Item_maxmin_subselect::Item_maxmin_subselect");
+  init(thd, select_lex, new select_max_min_finder_subselect(this, max));
+  max_columns= 1;
+  maybe_null= 1;
+  max_columns= 1;
+  DBUG_VOID_RETURN;
+}
+
 void Item_singlerow_subselect::reset()
 {
   null_value= 1;
@@ -499,38 +512,50 @@ Item_in_subselect::single_value_transformer(JOIN *join,
       (func == &Item_bool_func2::gt_creator ||
        func == &Item_bool_func2::lt_creator ||
        func == &Item_bool_func2::ge_creator ||
-       func == &Item_bool_func2::le_creator) &&
-      !select_lex->group_list.elements &&
-      !select_lex->with_sum_func)
+       func == &Item_bool_func2::le_creator))
   {
-    Item *item;
-    subs_type type= substype();
-    if (func == &Item_bool_func2::le_creator ||
-	func == &Item_bool_func2::lt_creator)
+    Item *subs;
+    if (!select_lex->group_list.elements &&
+	!select_lex->with_sum_func)
     {
-      /*
-	(ALL && (> || =>)) || (ANY && (< || =<))
-	for ALL condition is inverted
-      */
-      item= new Item_sum_max(*select_lex->ref_pointer_array);
+      Item *item;
+      subs_type type= substype();
+      if (func == &Item_bool_func2::le_creator ||
+	  func == &Item_bool_func2::lt_creator)
+      {
+	/*
+	  (ALL && (> || =>)) || (ANY && (< || =<))
+	  for ALL condition is inverted
+	*/
+	item= new Item_sum_max(*select_lex->ref_pointer_array);
+      }
+      else
+      {
+	/*
+	  (ALL && (< || =<)) || (ANY && (> || =>))
+	  for ALL condition is inverted
+	*/
+	item= new Item_sum_min(*select_lex->ref_pointer_array);
+      }
+      *select_lex->ref_pointer_array= item;
+      select_lex->item_list.empty();
+      select_lex->item_list.push_back(item);
+    
+      if (item->fix_fields(thd, join->tables_list, &item))
+      {
+	DBUG_RETURN(ERROR);
+      }
+      subs= new Item_singlerow_subselect(thd, select_lex);
     }
     else
     {
-      /*
-	(ALL && (< || =<)) || (ANY && (> || =>))
-	for ALL condition is inverted
-      */
-      item= new Item_sum_min(*select_lex->ref_pointer_array);
-    }
-    *select_lex->ref_pointer_array= item;
-    select_lex->item_list.empty();
-    select_lex->item_list.push_back(item);
-
-    if (item->fix_fields(thd, join->tables_list, &item))
-    {
-      DBUG_RETURN(ERROR);
+      // remove LIMIT placed  by ALL/ANY subquery
+      select_lex->master_unit()->global_parameters->select_limit=
+	HA_POS_ERROR;
+      subs= new Item_maxmin_subselect(thd, select_lex,
+				      (func == &Item_bool_func2::le_creator ||
+				       func == &Item_bool_func2::lt_creator));
     }
-
     // left expression belong to outer select
     SELECT_LEX *current= thd->lex.current_select, *up;
     thd->lex.current_select= up= current->return_after_parsing();
@@ -540,8 +565,7 @@ Item_in_subselect::single_value_transformer(JOIN *join,
       DBUG_RETURN(ERROR);
     }
     thd->lex.current_select= current;
-    substitution= (*func)(left_expr,
-			  new Item_singlerow_subselect(thd, select_lex));
+    substitution= (*func)(left_expr, subs);
     DBUG_RETURN(OK);
   }
 

sql/item_subselect.h

@@ -130,6 +130,7 @@ class Item_singlerow_subselect :public Item_subselect
     max_length= item->max_length;
     decimals= item->decimals;
   }
+  Item_singlerow_subselect() :Item_subselect(), value(0), row (0) {}
 
   subs_type substype() { return SINGLEROW_SUBS; }
 
@@ -153,6 +154,15 @@ class Item_singlerow_subselect :public Item_subselect
   friend class select_singlerow_subselect;
 };
 
+/* used in static ALL/ANY optimisation */
+class Item_maxmin_subselect: public Item_singlerow_subselect
+{
+public:
+  Item_maxmin_subselect(THD *thd, st_select_lex *select_lex, bool max);
+  Item_maxmin_subselect(Item_maxmin_subselect *item)
+    :Item_singlerow_subselect(item) {}
+};
+
 /* exists subselect */
 
 class Item_exists_subselect :public Item_subselect

sql/sql_class.cc

@@ -986,6 +986,95 @@ bool select_singlerow_subselect::send_data(List<Item> &items)
   DBUG_RETURN(0);
 }
 
+bool select_max_min_finder_subselect::send_data(List<Item> &items)
+{
+  DBUG_ENTER("select_max_min_finder_subselect::send_data");
+  Item_singlerow_subselect *it= (Item_singlerow_subselect *)item;
+  List_iterator_fast<Item> li(items);
+  Item *val_item= li++;
+  if (it->assigned())
+  {
+    cache->store(val_item);
+    if ((this->*op)())
+      it->store(0, cache);
+  }
+  else
+  {
+    if (!cache)
+    {
+      cache= Item_cache::get_cache(val_item->result_type());
+      switch (val_item->result_type())
+      {
+      case REAL_RESULT:
+	op= &select_max_min_finder_subselect::cmp_real;
+	break;
+      case INT_RESULT:
+	op= &select_max_min_finder_subselect::cmp_int;
+	break;
+      case STRING_RESULT:
+	op= &select_max_min_finder_subselect::cmp_str;
+	break;
+      case ROW_RESULT:
+        // This case should never be choosen
+	DBUG_ASSERT(0);
+	op= 0;
+      }
+    }
+    cache->store(val_item);
+    it->store(0, cache);
+  }
+  it->assigned(1);
+  DBUG_RETURN(0);
+}
+
+bool select_max_min_finder_subselect::cmp_real()
+{
+  Item *maxmin= ((Item_singlerow_subselect *)item)->el(0);
+  double val1= cache->val(), val2= maxmin->val();
+  if (fmax)
+    return (cache->null_value && !maxmin->null_value) ||
+      (!cache->null_value && !maxmin->null_value &&
+       val1 > val2);
+  else
+    return (maxmin->null_value && !cache->null_value) ||
+      (!cache->null_value && !maxmin->null_value &&
+       val1 < val2);
+}
+
+bool select_max_min_finder_subselect::cmp_int()
+{
+  Item *maxmin= ((Item_singlerow_subselect *)item)->el(0);
+  longlong val1= cache->val_int(), val2= maxmin->val_int();
+  if (fmax)
+    return (cache->null_value && !maxmin->null_value) ||
+      (!cache->null_value && !maxmin->null_value &&
+       val1 > val2);
+  else
+    return (maxmin->null_value && !cache->null_value) ||
+      (!cache->null_value && !maxmin->null_value &&
+       val1 < val2);
+}
+
+bool select_max_min_finder_subselect::cmp_str()
+{
+  String *val1, *val2, buf1, buf2;
+  Item *maxmin= ((Item_singlerow_subselect *)item)->el(0);
+  /*
+    as far as both operand is Item_cache buf1 & buf2 will not be used,
+    but added for safety
+  */
+  val1= cache->val_str(&buf1);
+  val2= maxmin->val_str(&buf1);
+  if (fmax)
+    return (cache->null_value && !maxmin->null_value) ||
+      (!cache->null_value && !maxmin->null_value &&
+       sortcmp(val1, val2, cache->collation.collation) > 0) ;
+  else
+    return (maxmin->null_value && !cache->null_value) ||
+      (!cache->null_value && !maxmin->null_value &&
+       sortcmp(val1, val2, cache->collation.collation) < 0);
+}
+
 bool select_exists_subselect::send_data(List<Item> &items)
 {
   DBUG_ENTER("select_exists_subselect::send_data");

sql/sql_class.h

@@ -920,6 +920,22 @@ class select_singlerow_subselect :public select_subselect
   bool send_data(List<Item> &items);
 };
 
+/* used in independent ALL/ANY optimisation */
+class select_max_min_finder_subselect :public select_subselect
+{
+  Item_cache *cache;
+  bool (select_max_min_finder_subselect::*op)();
+  bool fmax;
+public:
+  select_max_min_finder_subselect(Item_subselect *item, bool mx)
+    :select_subselect(item), cache(0), fmax(mx)
+  {}
+  bool send_data(List<Item> &items);
+  bool cmp_real();
+  bool cmp_int();
+  bool cmp_str();
+};
+
 /* EXISTS subselect interface class */
 class select_exists_subselect :public select_subselect
 {