Better bugfix for "HAVING when refering to RAND()" (Bug #8216)

Ensure that references in HAVING, ORDER BY or GROUP BY are calculated after fields in SELECT.
This will ensure that any reference to these has a valid value.
Generalized the code for split_sum_func()

.bzrignore

@@ -1007,3 +1007,4 @@ tests/mysql_client_test
 libmysqld/examples/mysql_client_test.c
 libmysqld/examples/mysql_client_test_embedded
 libmysqld/examples/mysqltest_embedded
+support-files/ndb-config-2-node.ini

mysql-test/r/group_by.result

@@ -655,9 +655,9 @@ insert into t1 (a,b) values (1,2),(1,3),(2,5);
 select a, 0.1*0+1 r2, sum(1) r1 from t1 where a = 1 group  by a having r1>1 and r2=1;
 a	r2	r1
 1	1.0	2
-select a, rand()*0+1 r2, sum(1) r1 from t1 where a = 1 group  by a having r1>1 and r2=1;
+select a, round(rand(100)*10) r2, sum(1) r1 from t1 where a = 1 group  by a having r1>1 and r2<=2;
 a	r2	r1
-1	1	2
+1	2	2
 select a,sum(b) from t1 where a=1 group by c;
 a	sum(b)
 1	5

mysql-test/r/user_var.result

@@ -109,8 +109,8 @@ select @a:=0;
 select @a, @a:=@a+count(*), count(*), @a from t1 group by i;
 @a	@a:=@a+count(*)	count(*)	@a
 0	1	1	0
-0	3	2	0
-0	6	3	0
+0	2	2	0
+0	3	3	0
 select @a:=0;
 @a:=0
 0

mysql-test/t/group_by.test

@@ -480,7 +480,8 @@ drop table t1;
 create table t1 (a integer, b integer, c integer);
 insert into t1 (a,b) values (1,2),(1,3),(2,5);
 select a, 0.1*0+1 r2, sum(1) r1 from t1 where a = 1 group  by a having r1>1 and r2=1;
-select a, rand()*0+1 r2, sum(1) r1 from t1 where a = 1 group  by a having r1>1 and r2=1;
+# rand(100)*10 will be < 2 only for the first row (of 6)
+select a, round(rand(100)*10) r2, sum(1) r1 from t1 where a = 1 group  by a having r1>1 and r2<=2;
 select a,sum(b) from t1 where a=1 group by c;
 select a*sum(b) from t1 where a=1 group by c;
 select sum(a)*sum(b) from t1 where a=1 group by c;

sql/item.cc

@@ -295,6 +295,58 @@ CHARSET_INFO *Item::default_charset()
 }
 
 
+/*
+  Move SUM items out from item tree and replace with reference
+
+  SYNOPSIS
+    split_sum_func2()
+    thd			Thread handler
+    ref_pointer_array	Pointer to array of reference fields
+    fields		All fields in select
+    ref			Pointer to item
+
+  NOTES
+   This is from split_sum_func2() for items that should be split
+
+   All found SUM items are added FIRST in the fields list and
+   we replace the item with a reference.
+
+   thd->fatal_error() may be called if we are out of memory
+*/
+
+
+void Item::split_sum_func2(THD *thd, Item **ref_pointer_array,
+                           List<Item> &fields, Item **ref)
+{
+  if (type() != SUM_FUNC_ITEM && with_sum_func)
+  {
+    /* Will split complicated items and ignore simple ones */
+    split_sum_func(thd, ref_pointer_array, fields);
+  }
+  else if ((type() == SUM_FUNC_ITEM ||
+            (used_tables() & ~PARAM_TABLE_BIT)) &&
+           type() != REF_ITEM)
+  {
+    /*
+      Replace item with a reference so that we can easily calculate
+      it (in case of sum functions) or copy it (in case of fields)
+
+      The test above is to ensure we don't do a reference for things
+      that are constants (PARAM_TABLE_BIT is in effect a constant)
+      or already referenced (for example an item in HAVING)
+    */
+    uint el= fields.elements;
+    Item *new_item;    
+    ref_pointer_array[el]= this;
+    if (!(new_item= new Item_ref(ref_pointer_array + el, 0, name)))
+      return;                                   // fatal_error is set
+    fields.push_front(this);
+    ref_pointer_array[el]= this;
+    thd->change_item_tree(ref, new_item);
+  }
+}
+
+
 /*
    Aggregate two collations together taking
    into account their coercibility (aka derivation):

sql/item.h

@@ -262,6 +262,9 @@ public:
   virtual void update_used_tables() {}
   virtual void split_sum_func(THD *thd, Item **ref_pointer_array,
                               List<Item> &fields) {}
+  /* Called for items that really have to be split */
+  void split_sum_func2(THD *thd, Item **ref_pointer_array, List<Item> &fields,
+                       Item **ref);
   virtual bool get_date(TIME *ltime,uint fuzzydate);
   virtual bool get_time(TIME *ltime);
   virtual bool get_date_result(TIME *ltime,uint fuzzydate)

sql/item_cmpfunc.cc

@@ -1969,10 +1969,10 @@ bool Item_cond::walk(Item_processor processor, byte *arg)
   Move SUM items out from item tree and replace with reference
 
   SYNOPSIS
-  split_sum_func()
-  thd			Thread handler
-  ref_pointer_array	Pointer to array of reference fields
-  fields		All fields in select
+    split_sum_func()
+    thd			Thread handler
+    ref_pointer_array	Pointer to array of reference fields
+    fields		All fields in select
 
   NOTES
    This function is run on all expression (SELECT list, WHERE, HAVING etc)
@@ -1982,16 +1982,6 @@ bool Item_cond::walk(Item_processor processor, byte *arg)
    so that we can easily find and calculate them.
    (Calculation done by update_sum_func() and copy_sum_funcs() in
    sql_select.cc)
-
-   All found SUM items are added FIRST in the fields list and
-   we replace the item with a reference.
-
-   We also replace all functions without side effects (like RAND() or UDF's)
-   that uses columns as arguments.
-   For functions with side effects, we just remember any fields referred
-   by the function to ensure that we get a copy of the field value for the
-   first accepted row. This ensures that we can do things like
-   SELECT a*SUM(b) FROM t1 WHERE a=1
 */
 
 void Item_cond::split_sum_func(THD *thd, Item **ref_pointer_array,
@@ -1999,38 +1989,8 @@ void Item_cond::split_sum_func(THD *thd, Item **ref_pointer_array,
 {
   List_iterator<Item> li(list);
   Item *item;
-  used_tables_cache=0;
-  const_item_cache=0;
-  while ((item=li++))
-  {
-    /* with_sum_func is set for items that contains a SUM expression */
-    if (item->type() != SUM_FUNC_ITEM &&
-        (item->with_sum_func ||
-         (item->used_tables() & PSEUDO_TABLE_BITS)))
-      item->split_sum_func(thd, ref_pointer_array, fields);
-    else if (item->type() == SUM_FUNC_ITEM ||
-             (item->used_tables() && item->type() != REF_ITEM))
-    {
-      /*
-        Replace item with a reference so that we can easily calculate
-        it (in case of sum functions) or copy it (in case of fields)
-
-        The test above is to ensure we don't do a reference for things
-        that are constants or are not yet calculated as in:
-        SELECT RAND() as r1, SUM(a) as r2 FROM t1 HAVING r1 > 1 AND r2 > 0
-      */
-      Item **ref= li.ref();
-      uint el= fields.elements;
-      ref_pointer_array[el]= item;
-      Item *new_item= new Item_ref(ref_pointer_array + el, 0, item->name);
-      fields.push_front(item);
-      ref_pointer_array[el]= item;
-      thd->change_item_tree(ref, new_item);
-    }
-    item->update_used_tables();
-    used_tables_cache|=item->used_tables();
-    const_item_cache&=item->const_item();
-  }
+  while ((item= li++))
+    item->split_sum_func2(thd, ref_pointer_array, fields, li.ref());
 }
 
 

sql/item_func.cc

@@ -352,28 +352,14 @@ bool Item_func::walk (Item_processor processor, byte *argument)
 }
 
 
+/* See comments in Item_cmp_func::split_sum_func() */
+
 void Item_func::split_sum_func(THD *thd, Item **ref_pointer_array,
                                List<Item> &fields)
 {
   Item **arg, **arg_end;
   for (arg= args, arg_end= args+arg_count; arg != arg_end ; arg++)
-  {
-    Item *item=* arg;
-    if (item->type() != SUM_FUNC_ITEM &&
-        (item->with_sum_func ||
-         (item->used_tables() & PSEUDO_TABLE_BITS)))
-      item->split_sum_func(thd, ref_pointer_array, fields);
-    else if (item->type() == SUM_FUNC_ITEM ||
-             (item->used_tables() && item->type() != REF_ITEM))
-    {
-      uint el= fields.elements;
-      ref_pointer_array[el]= item;
-      Item *new_item= new Item_ref(ref_pointer_array + el, 0, item->name);
-      fields.push_front(item);
-      ref_pointer_array[el]= item;
-      thd->change_item_tree(arg, new_item);
-    }
-  }
+    (*arg)->split_sum_func2(thd, ref_pointer_array, fields, arg);
 }
 
 

sql/item_row.cc

@@ -90,25 +90,10 @@ void Item_row::split_sum_func(THD *thd, Item **ref_pointer_array,
 {
   Item **arg, **arg_end;
   for (arg= items, arg_end= items+arg_count; arg != arg_end ; arg++)
-  {
-    Item *item= *arg;
-    if (item->type() != SUM_FUNC_ITEM &&
-        (item->with_sum_func ||
-         (item->used_tables() & PSEUDO_TABLE_BITS)))
-      item->split_sum_func(thd, ref_pointer_array, fields);
-    else if (item->type() == SUM_FUNC_ITEM ||
-             (item->used_tables() && item->type() != REF_ITEM))
-    {
-      uint el= fields.elements;
-      ref_pointer_array[el]=*arg;
-      Item *new_item= new Item_ref(ref_pointer_array + el, 0, (*arg)->name);
-      fields.push_front(*arg);
-      ref_pointer_array[el]= *arg;
-      thd->change_item_tree(arg, new_item);
-    }
-  }
+    (*arg)->split_sum_func2(thd, ref_pointer_array, fields, arg);
 }
 
+
 void Item_row::update_used_tables()
 {
   used_tables_cache= 0;

sql/item_strfunc.cc

@@ -1758,20 +1758,7 @@ String *Item_func_elt::val_str(String *str)
 void Item_func_make_set::split_sum_func(THD *thd, Item **ref_pointer_array,
 					List<Item> &fields)
 {
-  if (item->type() != SUM_FUNC_ITEM &&
-      (item->with_sum_func ||
-       (item->used_tables() & PSEUDO_TABLE_BITS)))
-    item->split_sum_func(thd, ref_pointer_array, fields);
-  else if (item->type() == SUM_FUNC_ITEM ||
-           (item->used_tables() && item->type() != REF_ITEM))
-  {
-    uint el= fields.elements;
-    ref_pointer_array[el]=item;
-    Item *new_item= new Item_ref(ref_pointer_array + el, 0, item->name);
-    fields.push_front(item);
-    ref_pointer_array[el]= item;
-    thd->change_item_tree(&item, new_item);
-  }
+  item->split_sum_func2(thd, ref_pointer_array, fields, &item);
   Item_str_func::split_sum_func(thd, ref_pointer_array, fields);
 }
 

sql/sql_list.h

@@ -134,6 +134,15 @@ public:
     if (!--elements)
       last= &first;
   }
+  inline void concat(base_list *list)
+  {
+    if (!list->is_empty())
+    {
+      *last= list->first;
+      last= list->last;
+      elements+= list->elements;
+    }
+  }
   inline void *pop(void)
   {
     if (first == &end_of_list) return 0;

sql/sql_select.cc

@@ -8502,6 +8502,7 @@ setup_copy_fields(THD *thd, TMP_TABLE_PARAM *param,
   res_selected_fields.empty();
   res_all_fields.empty();
   List_iterator_fast<Item> itr(res_all_fields);
+  List<Item> extra_funcs;
   uint i, border= all_fields.elements - elements;
   DBUG_ENTER("setup_copy_fields");
 
@@ -8563,7 +8564,12 @@ setup_copy_fields(THD *thd, TMP_TABLE_PARAM *param,
       */
       if (!(pos=new Item_copy_string(pos)))
 	goto err;
-      if (param->copy_funcs.push_back(pos))
+      if (i < border)                           // HAVING, ORDER and GROUP BY
+      {
+        if (extra_funcs.push_back(pos))
+          goto err;
+      }
+      else if (param->copy_funcs.push_back(pos))
 	goto err;
     }
     res_all_fields.push_back(pos);
@@ -8575,6 +8581,12 @@ setup_copy_fields(THD *thd, TMP_TABLE_PARAM *param,
   for (i= 0; i < border; i++)
     itr++;
   itr.sublist(res_selected_fields, elements);
+  /*
+    Put elements from HAVING, ORDER BY and GROUP BY last to ensure that any
+    reference used in these will resolve to a item that is already calculated
+  */
+  param->copy_funcs.concat(&extra_funcs);
+
   DBUG_RETURN(0);
 
  err: