optimized IN with Rows (SCRUM)

NULL with row (IN)
cardinality error of row inside row
new[]
fixed layout

mysql-test/r/row.result

@@ -4,6 +4,35 @@ row(1,2,3) IN (row(3,2,3), row(1,2,3), row(1,3,3))
 select row(10,2,3) IN (row(3,2,3), row(1,2,3), row(1,3,3));
 row(10,2,3) IN (row(3,2,3), row(1,2,3), row(1,3,3))
 0
+select row(1,2,3) IN (row(3,NULL,3), row(1,2,3), row(1,3,3));
+row(1,2,3) IN (row(3,NULL,3), row(1,2,3), row(1,3,3))
+1
+select row(10,2,3) IN (row(3,NULL,3), row(1,2,3), row(1,3,3));
+row(10,2,3) IN (row(3,NULL,3), row(1,2,3), row(1,3,3))
+NULL
+select row('a',1.5,3) IN (row(1,2,3), row('a',1.5,3), row('a','a','a'));
+row('a',1.5,3) IN (row(1,2,3), row('a',1.5,3), row('a','a','a'))
+1
+select row('a',0,3) IN (row(3,2,3), row('a','0','3'), row(1,3,3));
+row('a',0,3) IN (row(3,2,3), row('a','0','3'), row(1,3,3))
+1
+select row('a',0,3) IN (row(3,2,3), row('a','a','3'), row(1,3,3));
+row('a',0,3) IN (row(3,2,3), row('a','a','3'), row(1,3,3))
+1
+select row('a',1.5,3) IN (row(3,NULL,3), row('a',1.5,3), row(1,3,3));
+row('a',1.5,3) IN (row(3,NULL,3), row('a',1.5,3), row(1,3,3))
+1
+select row('b',1.5,3) IN (row(3,NULL,3), row('a',1.5,3), row(1,3,3));
+row('b',1.5,3) IN (row(3,NULL,3), row('a',1.5,3), row(1,3,3))
+NULL
+select row(1,2,row(3,4)) IN (row(3,2,row(3,4)), row(1,2,row(3,4)));
+row(1,2,row(3,4)) IN (row(3,2,row(3,4)), row(1,2,row(3,4)))
+1
+select row(1,2,row(3,4)) IN (row(3,2,row(3,4)), row(1,2,4));
+Cardinality error (more/less than 2 columns)
+select row(1,2,row(3,4)) IN (row(3,2,row(3,4)), row(1,2,row(3,NULL)));
+row(1,2,row(3,4)) IN (row(3,2,row(3,4)), row(1,2,row(3,NULL)))
+NULL
 SELECT ROW(1,2,3)=ROW(1,2,3);
 ROW(1,2,3)=ROW(1,2,3)
 1
@@ -42,9 +71,20 @@ ROW('test',2,3.33)=ROW('test',2,3.33)
 1
 SELECT ROW('test',2,3.33)=ROW('test',2,3.33,4);
 Cardinality error (more/less than 3 columns)
+SELECT ROW('test',2,ROW(3,33))=ROW('test',2,ROW(3,33));
+ROW('test',2,ROW(3,33))=ROW('test',2,ROW(3,33))
+1
+SELECT ROW('test',2,ROW(3,33))=ROW('test',2,ROW(3,3));
+ROW('test',2,ROW(3,33))=ROW('test',2,ROW(3,3))
+0
+SELECT ROW('test',2,ROW(3,33))=ROW('test',2,ROW(3,NULL));
+ROW('test',2,ROW(3,33))=ROW('test',2,ROW(3,NULL))
+NULL
+SELECT ROW('test',2,ROW(3,33))=ROW('test',2,4);
+Cardinality error (more/less than 2 columns)
 drop table if exists t1;
 create table t1 ( a int, b int, c int);
-insert into t1 values (1,2,3), (2,3,1), (3,2,1);
+insert into t1 values (1,2,3), (2,3,1), (3,2,1), (1,2,NULL);
 select * from t1 where ROW(1,2,3)=ROW(a,b,c);
 a	b	c
 1	2	3
@@ -54,16 +94,30 @@ select * from t1 where ROW(1,2,3)<ROW(a,b,c);
 a	b	c
 2	3	1
 3	2	1
-select * from t1 where ROW(a,2,3) IN(row(1,b,c), row(2,3,1));
-a	b	c
-1	2	3
-select * from t1 where ROW(c,2,3) IN(row(1,b,a), row(2,3,1));
-a	b	c
-3	2	1
-select * from t1 where ROW(a,b,c) IN(row(1,2,3), row(3,2,1));
-a	b	c
-1	2	3
-3	2	1
+select ROW(a,2,3) IN(row(1,b,c), row(2,3,1)) from t1;
+ROW(a,2,3) IN(row(1,b,c), row(2,3,1))
+1
+0
+0
+NULL
+select ROW(c,2,3) IN(row(1,b,a), row(2,3,1)) from t1;
+ROW(c,2,3) IN(row(1,b,a), row(2,3,1))
+0
+0
+1
+NULL
+select ROW(a,b,c) IN(row(1,2,3), row(3,2,1)) from t1;
+ROW(a,b,c) IN(row(1,2,3), row(3,2,1))
+1
+0
+1
+NULL
+select ROW(1,2,3) IN(row(a,b,c), row(1,2,3)) from t1;
+ROW(1,2,3) IN(row(a,b,c), row(1,2,3))
+1
+1
+1
+1
 drop table t1;
 select ROW(1,1);
 Cardinality error (more/less than 1 columns)

mysql-test/t/row.test

 select row(1,2,3) IN (row(3,2,3), row(1,2,3), row(1,3,3));
 select row(10,2,3) IN (row(3,2,3), row(1,2,3), row(1,3,3));
+select row(1,2,3) IN (row(3,NULL,3), row(1,2,3), row(1,3,3));
+select row(10,2,3) IN (row(3,NULL,3), row(1,2,3), row(1,3,3));
+select row('a',1.5,3) IN (row(1,2,3), row('a',1.5,3), row('a','a','a'));
+select row('a',0,3) IN (row(3,2,3), row('a','0','3'), row(1,3,3));
+select row('a',0,3) IN (row(3,2,3), row('a','a','3'), row(1,3,3));
+select row('a',1.5,3) IN (row(3,NULL,3), row('a',1.5,3), row(1,3,3));
+select row('b',1.5,3) IN (row(3,NULL,3), row('a',1.5,3), row(1,3,3));
+select row(1,2,row(3,4)) IN (row(3,2,row(3,4)), row(1,2,row(3,4)));
+-- error 1239
+select row(1,2,row(3,4)) IN (row(3,2,row(3,4)), row(1,2,4));
+select row(1,2,row(3,4)) IN (row(3,2,row(3,4)), row(1,2,row(3,NULL)));
 
 SELECT ROW(1,2,3)=ROW(1,2,3);
 SELECT ROW(2,2,3)=ROW(1+1,2,3);
@@ -15,15 +26,21 @@ SELECT ROW(1,2,ROW(3,4,5))=ROW(1,2,ROW(3,4,5));
 SELECT ROW('test',2,3.33)=ROW('test',2,3.33);
 -- error 1239
 SELECT ROW('test',2,3.33)=ROW('test',2,3.33,4);
+SELECT ROW('test',2,ROW(3,33))=ROW('test',2,ROW(3,33));
+SELECT ROW('test',2,ROW(3,33))=ROW('test',2,ROW(3,3));
+SELECT ROW('test',2,ROW(3,33))=ROW('test',2,ROW(3,NULL));
+-- error 1239
+SELECT ROW('test',2,ROW(3,33))=ROW('test',2,4);
 drop table if exists t1;
 create table t1 ( a int, b int, c int);
-insert into t1 values (1,2,3), (2,3,1), (3,2,1);
+insert into t1 values (1,2,3), (2,3,1), (3,2,1), (1,2,NULL);
 select * from t1 where ROW(1,2,3)=ROW(a,b,c);
 select * from t1 where ROW(0,2,3)=ROW(a,b,c);
 select * from t1 where ROW(1,2,3)<ROW(a,b,c);
-select * from t1 where ROW(a,2,3) IN(row(1,b,c), row(2,3,1));
-select * from t1 where ROW(c,2,3) IN(row(1,b,a), row(2,3,1));
-select * from t1 where ROW(a,b,c) IN(row(1,2,3), row(3,2,1));
+select ROW(a,2,3) IN(row(1,b,c), row(2,3,1)) from t1;
+select ROW(c,2,3) IN(row(1,b,a), row(2,3,1)) from t1;
+select ROW(a,b,c) IN(row(1,2,3), row(3,2,1)) from t1;
+select ROW(1,2,3) IN(row(a,b,c), row(1,2,3)) from t1;
 drop table t1;
 
 -- error 1239

sql/item_cmpfunc.cc

@@ -144,7 +144,14 @@ int Arg_comparator::set_compare_func(Item_bool_func2 *item, Item_result type)
     }
     if ((comparators= (Arg_comparator *) sql_alloc(sizeof(Arg_comparator)*n)))
       for (uint i=0; i < n; i++)
+      {
+	if ((*a)->cols() != (*a)->cols())
+	{
+	  my_error(ER_CARDINALITY_COL, MYF(0), (*a)->cols());
+	  return 1;
+	}
 	comparators[i].set_cmp_func(owner, (*a)->addr(i), (*b)->addr(i));
+      }
     else
     {
       my_message(ER_OUT_OF_RESOURCES, ER(ER_OUT_OF_RESOURCES), MYF(0));
@@ -1033,6 +1040,11 @@ static int cmp_double(double *a,double *b)
   return *a < *b ? -1 : *a == *b ? 0 : 1;
 }
 
+static int cmp_row(cmp_item_row* a, cmp_item_row* b)
+{
+  return a->compare(b);
+}
+
 int in_vector::find(Item *item)
 {
   byte *result=get_value(item);
@@ -1055,7 +1067,6 @@ int in_vector::find(Item *item)
   return (int) ((*compare)(base+start*size,result) == 0);
 }
 
-
 in_string::in_string(uint elements,qsort_cmp cmp_func)
   :in_vector(elements,sizeof(String),cmp_func),tmp(buff,sizeof(buff),default_charset_info)
 {}
@@ -1082,6 +1093,29 @@ byte *in_string::get_value(Item *item)
   return (byte*) item->val_str(&tmp);
 }
 
+in_row::in_row(uint elements, Item * item)
+{
+  DBUG_ENTER("in_row::in_row");
+  base= (char*) new cmp_item_row[elements];
+  size= sizeof(cmp_item_row);
+  compare= (qsort_cmp) cmp_row;
+  tmp.store_value(item);
+  DBUG_VOID_RETURN;
+}
+
+byte *in_row::get_value(Item *item)
+{
+  tmp.store_value(item);
+  return (byte *)&tmp;
+}
+
+void in_row::set(uint pos, Item *item)
+{
+  DBUG_ENTER("in_row::set");
+  DBUG_PRINT("enter", ("pos %u item 0x%lx", pos, (ulong) item));
+  ((cmp_item_row*) base)[pos].store_value_by_template(&tmp, item);
+  DBUG_VOID_RETURN;
+}
 
 in_longlong::in_longlong(uint elements)
   :in_vector(elements,sizeof(longlong),(qsort_cmp) cmp_longlong)
@@ -1094,13 +1128,12 @@ void in_longlong::set(uint pos,Item *item)
 
 byte *in_longlong::get_value(Item *item)
 {
-  tmp=item->val_int();
+  tmp= item->val_int();
   if (item->null_value)
-    return 0;					/* purecov: inspected */
+    return 0;
   return (byte*) &tmp;
 }
 
-
 in_double::in_double(uint elements)
   :in_vector(elements,sizeof(double),(qsort_cmp) cmp_double)
 {}
@@ -1112,7 +1145,7 @@ void in_double::set(uint pos,Item *item)
 
 byte *in_double::get_value(Item *item)
 {
-  tmp=item->val();
+  tmp= item->val();
   if (item->null_value)
     return 0;					/* purecov: inspected */
   return (byte*) &tmp;
@@ -1140,14 +1173,77 @@ cmp_item* cmp_item::get_comparator (Item *item)
   return 0; // to satisfy compiler :)
 }
 
+cmp_item* cmp_item_sort_string::make_same()
+{
+  return new cmp_item_sort_string_in_static();
+}
+
+cmp_item* cmp_item_binary_string::make_same()
+{
+  return new cmp_item_binary_string_in_static();
+}
+
+cmp_item* cmp_item_int::make_same()
+{
+  return new cmp_item_int();
+}
+
+cmp_item* cmp_item_real::make_same()
+{
+  return new cmp_item_real();
+}
+
+cmp_item* cmp_item_row::make_same()
+{
+  return new cmp_item_row();
+}
+
 void cmp_item_row::store_value(Item *item)
 {
   n= item->cols();
   if ((comparators= (cmp_item **) sql_alloc(sizeof(cmp_item *)*n)))
   {
+    item->null_value= 0;
     for (uint i=0; i < n; i++)
       if ((comparators[i]= cmp_item::get_comparator(item->el(i))))
+      {
 	comparators[i]->store_value(item->el(i));
+	item->null_value|= item->el(i)->null_value;
+      }
+      else
+      {
+	my_message(ER_OUT_OF_RESOURCES, ER(ER_OUT_OF_RESOURCES), MYF(0));
+	current_thd->fatal_error= 1;
+	return;
+      }	  
+  }
+  else
+  {
+    my_message(ER_OUT_OF_RESOURCES, ER(ER_OUT_OF_RESOURCES), MYF(0));
+    current_thd->fatal_error= 1;
+    return;
+  }
+}
+
+void cmp_item_row::store_value_by_template(cmp_item *t, Item *item)
+{
+  cmp_item_row *tmpl= (cmp_item_row*) t;
+  if (tmpl->n != item->cols())
+  {
+    my_error(ER_CARDINALITY_COL, MYF(0), tmpl->n);
+    return;
+  }
+  n= tmpl->n;
+  if ((comparators= (cmp_item **) sql_alloc(sizeof(cmp_item *)*n)))
+  {
+    item->null_value= 0;
+    for (uint i=0; i < n; i++)
+      if ((comparators[i]= tmpl->comparators[i]->make_same()))
+      {
+	comparators[i]->store_value_by_template(tmpl->comparators[i],
+						item->el(i));
+	item->null_value|= item->el(i)->null_value;
+      }
       else
       {
 	my_message(ER_OUT_OF_RESOURCES, ER(ER_OUT_OF_RESOURCES), MYF(0));
@@ -1165,15 +1261,34 @@ void cmp_item_row::store_value(Item *item)
 
 int cmp_item_row::cmp(Item *arg)
 {
+  arg->null_value= 0;
+  if (arg->cols() != n)
+  {
+    my_error(ER_CARDINALITY_COL, MYF(0), n);
+    return 1;
+  }
   for(uint i=0; i < n; i++)
     if(comparators[i]->cmp(arg->el(i)))
+    {
+      arg->null_value|= arg->el(i)->null_value;
       return 1;
+    }
+  return 0;
+}
+
+int cmp_item_row::compare(cmp_item *c)
+{
+  int res;
+  cmp_item_row *cmp= (cmp_item_row *) c;
+  for(uint i=0; i < n; i++)
+    if((res= comparators[i]->compare(cmp->comparators[i])))
+      return res;
   return 0;
 }
 
 void Item_func_in::fix_length_and_dec()
 {
-  if (const_item() && item->result_type()!=ROW_RESULT)
+  if (const_item())
   {
     switch (item->result_type()) {
     case STRING_RESULT:
@@ -1189,8 +1304,7 @@ void Item_func_in::fix_length_and_dec()
       array= new in_double(arg_count);
       break;
     case ROW_RESULT:
-      // This case should never be choosen
-      DBUG_ASSERT(0);
+      array= new in_row(arg_count, item);
       break;
     }
     uint j=0;

sql/item_cmpfunc.h

@@ -17,6 +17,8 @@
 
 /* compare and test functions */
 
+#include "assert.h"
+
 #ifdef __GNUC__
 #pragma interface			/* gcc class implementation */
 #endif
@@ -375,6 +377,7 @@ class in_vector :public Sql_alloc
   uint count;
 public:
   uint used_count;
+  in_vector() {}
   in_vector(uint elements,uint element_length,qsort_cmp cmp_func)
     :base((char*) sql_calloc(elements*element_length)),
      size(element_length), compare(cmp_func), count(elements),
@@ -389,7 +392,6 @@ class in_vector :public Sql_alloc
   int find(Item *item);
 };
 
-
 class in_string :public in_vector
 {
   char buff[80];
@@ -401,7 +403,6 @@ class in_string :public in_vector
   byte *get_value(Item *item);
 };
 
-
 class in_longlong :public in_vector
 {
   longlong tmp;
@@ -411,7 +412,6 @@ class in_longlong :public in_vector
   byte *get_value(Item *item);
 };
 
-
 class in_double :public in_vector
 {
   double tmp;
@@ -421,7 +421,6 @@ class in_double :public in_vector
   byte *get_value(Item *item);
 };
 
-
 /*
 ** Classes for easy comparing of non const items
 */
@@ -431,30 +430,55 @@ class cmp_item :public Sql_alloc
 public:
   cmp_item() {}
   virtual ~cmp_item() {}
-  virtual void store_value(Item *item)=0;
-  virtual int cmp(Item *item)=0;
+  virtual void store_value(Item *item)= 0;
+  virtual int cmp(Item *item)= 0;
+  // for optimized IN with row
+  virtual int compare(cmp_item *item)= 0;
   static cmp_item* get_comparator(Item *);
+  virtual cmp_item *make_same()= 0;
+  virtual void store_value_by_template(cmp_item *tmpl, Item *item)
+  {
+    store_value(item);
+  }
 };
 
+class cmp_item_string :public cmp_item 
+{
+protected:
+  String *value_res;
+public:
+  friend class cmp_item_sort_string;
+  friend class cmp_item_binary_string;
+  friend class cmp_item_sort_string_in_static;
+  friend class cmp_item_binary_string_in_static;
+};
 
-class cmp_item_sort_string :public cmp_item {
- protected:
+class cmp_item_sort_string :public cmp_item_string
+{
+protected:
   char value_buff[80];
-  String value,*value_res;
+  String value;
 public:
-  cmp_item_sort_string() :value(value_buff,sizeof(value_buff),default_charset_info) {}
+  cmp_item_sort_string():
+    value(value_buff, sizeof(value_buff), default_charset_info) {}
   void store_value(Item *item)
   {
-      value_res=item->val_str(&value);
+    value_res= item->val_str(&value);
   }
   int cmp(Item *arg)
   {
     char buff[80];
-      String tmp(buff,sizeof(buff),default_charset_info),*res;
-      if (!(res=arg->val_str(&tmp)))
+    String tmp(buff, sizeof(buff), default_charset_info), *res;
+    if (!(res= arg->val_str(&tmp)))
       return 1;				/* Can't be right */
-      return sortcmp(value_res,res);
+    return sortcmp(value_res, res);
   }
+  int compare(cmp_item *c)
+  {
+    cmp_item_string *cmp= (cmp_item_string *)c;
+    return sortcmp(value_res, cmp->value_res);
+  } 
+  cmp_item *make_same();
 };
 
 class cmp_item_binary_string :public cmp_item_sort_string {
@@ -468,24 +492,34 @@ class cmp_item_binary_string :public cmp_item_sort_string {
       return 1;				/* Can't be right */
     return stringcmp(value_res,res);
   }
+  int compare(cmp_item *c)
+  {
+    cmp_item_string *cmp= (cmp_item_string *)c;
+    return stringcmp(value_res, cmp->value_res);
+  }
+  cmp_item *make_same();
 };
 
-
 class cmp_item_int :public cmp_item
 {
   longlong value;
 public:
   void store_value(Item *item)
   {
-      value=item->val_int();
+    value= item->val_int();
   }
   int cmp(Item *arg)
   {
     return value != arg->val_int();
   }
+  int compare(cmp_item *c)
+  {
+    cmp_item_int *cmp= (cmp_item_int *)c;
+    return (value < cmp->value) ? -1 : ((value == cmp->value) ? 0 : 1);
+  }
+  cmp_item *make_same();
 };
 
-
 class cmp_item_real :public cmp_item
 {
   double value;
@@ -498,6 +532,12 @@ class cmp_item_real :public cmp_item
   {
     return value != arg->val();
   }
+  int compare(cmp_item *c)
+  {
+    cmp_item_real *cmp= (cmp_item_real *)c;
+    return (value < cmp->value)? -1 : ((value == cmp->value) ? 0 : 1);
+  }
+  cmp_item *make_same();
 };
 
 class cmp_item_row :public cmp_item
@@ -505,8 +545,75 @@ class cmp_item_row :public cmp_item
   cmp_item **comparators;
   uint n;
 public:
+  cmp_item_row(): comparators(0), n(0) {}
+  ~cmp_item_row()
+  {
+    if(comparators)
+      for(uint i= 0; i < n; i++)
+	if (comparators[i])
+	  delete comparators[i];
+  }
   void store_value(Item *item);
   int cmp(Item *arg);
+  int compare(cmp_item *arg);
+  cmp_item *make_same();
+  void store_value_by_template(cmp_item *tmpl, Item *);
+};
+
+
+class in_row :public in_vector
+{
+  cmp_item_row tmp;
+public:
+  in_row(uint elements, Item *);
+  void set(uint pos,Item *item);
+  byte *get_value(Item *item);
+};
+
+/* 
+   cmp_item for optimized IN with row (right part string, which never
+   be changed)
+*/
+
+class cmp_item_sort_string_in_static :public cmp_item_string
+{
+ protected:
+  String value;
+public:
+  cmp_item_sort_string_in_static() {}
+  void store_value(Item *item)
+  {
+    value_res= item->val_str(&value);
+  }
+  int cmp(Item *item)
+  {
+    // Should never be called
+    DBUG_ASSERT(0);
+    return 1;
+  }
+  int compare(cmp_item *c)
+  {
+    cmp_item_string *cmp= (cmp_item_string *)c;
+    return sortcmp(value_res, cmp->value_res);
+  }
+  cmp_item * make_same()
+  {
+    return new cmp_item_sort_string_in_static();
+  }
+};
+
+class cmp_item_binary_string_in_static :public cmp_item_sort_string_in_static {
+public:
+  cmp_item_binary_string_in_static() {}
+  int compare(cmp_item *c)
+  {
+    cmp_item_string *cmp= (cmp_item_string *)c;
+    return stringcmp(value_res, cmp->value_res);
+  }
+  cmp_item * make_same()
+  {
+    return new cmp_item_binary_string_in_static();
+  }
 };
 
 class Item_func_in :public Item_int_func
@@ -546,8 +653,6 @@ class Item_func_in :public Item_int_func
   }
 };
 
-
-
 /* Functions used by where clause */
 
 class Item_func_isnull :public Item_bool_func

sql/sql_list.h

@@ -25,8 +25,16 @@
 class Sql_alloc
 {
 public:
-  static void *operator new(size_t size) {return (void*) sql_alloc((uint) size); }
+  static void *operator new(size_t size)
+  {
+    return (void*) sql_alloc((uint) size);
+  }
+  static void *operator new[](size_t size)
+  {
+    return (void*) sql_alloc((uint) size);
+  }
   static void operator delete(void *ptr, size_t size) {} /*lint -e715 */
+  static void operator delete[](void *ptr, size_t size) {}
 #ifdef HAVE_purify
   bool dummy;
   inline Sql_alloc() :dummy(0) {}