From: Date: September 12 2008 3:29pm
Subject: bzr commit into mysql-6.0-opt branch (sergefp:2686) Bug#38049
List-Archive: http://lists.mysql.com/commits/54002
X-Bug: 38049
Message-Id: <20080912132947.CFDE215A312@pslp2.localdomain>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7BIT

#At file:///home/spetrunia/dev/mysql-6.0-bug38049/

 2686 Sergey Petrunia	2008-09-12
      BUG#38049: incorrect rows estimations with references from preceding table
      - Fix condition in ReuseRangeEstimateForRef-3. The intended meaning of the changed
        part is: "And ref access candidate has key=const restrictions for every keypart that 
        was used by range access".
added:
  mysql-test/r/join_optimizer.result
  mysql-test/t/join_optimizer.test
modified:
  sql/sql_select.cc

per-file messages:
  mysql-test/r/join_optimizer.result
    BUG#38049: incorrect rows estimations with references from preceding table
    - Testcase
  mysql-test/t/join_optimizer.test
    BUG#38049: incorrect rows estimations with references from preceding table
    - Testcase
  sql/sql_select.cc
    BUG#38049: incorrect rows estimations with references from preceding table
    - Fix condition in ReuseRangeEstimateForRef-3. The intended meaning of the changed
      part is: "And ref access candidate has key=const restrictions for every keypart that 
      was used by range access".
=== added file 'mysql-test/r/join_optimizer.result'
--- a/mysql-test/r/join_optimizer.result	1970-01-01 00:00:00 +0000
+++ b/mysql-test/r/join_optimizer.result	2008-09-12 13:29:34 +0000
@@ -0,0 +1,37 @@
+drop table if exists t0,t1,t2,t3;
+#
+# BUG#38049 incorrect rows estimations with references from preceding table
+#
+create table t0 (a int);
+insert into t0 values (0),(1),(2),(3),(4),(5),(6),(7),(8),(9);
+create table t1 (a varchar(32));
+insert into t1 values ('owner'),('requester'),('admincc'),('cc');
+CREATE TABLE t2 (
+id int(11) NOT NULL,
+type varchar(32) default NULL,
+PRIMARY KEY  (id)
+);
+insert into t2 values (1,'owner'), (2,'admincc');
+CREATE TABLE t3 (
+id int(11) NOT NULL,
+domain varchar(32) default NULL,
+type varchar(32) default NULL,
+PRIMARY KEY  (id)
+);
+set @domain='system';
+set @pk=0;
+INSERT INTO t3 select @pk:=@pk+1, 'system', t1.a from t1;
+INSERT INTO t3 select @pk:=@pk+1, 'queue', t1.a from t1, t0 where t0.a<3;
+INSERT INTO t3 select @pk:=@pk+1, 'ticket', t1.a from t1, t0 A, t0 B, t0 C;
+CREATE INDEX groups_d ON t3(domain);
+CREATE INDEX groups_t ON t3(type);
+CREATE INDEX groups_td ON t3(type, domain);
+CREATE INDEX groups_dt ON t3(domain, type);
+For table g this must use ref(groups_dt) and #rows should be around 15 and not 335:
+explain 
+SELECT STRAIGHT_JOIN g.id FROM t2 a, t3 g USE INDEX(groups_dt) 
+WHERE g.domain = 'queue' AND g.type = a.type;
+id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
+1	SIMPLE	a	ALL	NULL	NULL	NULL	NULL	2	
+1	SIMPLE	g	ref	groups_dt	groups_dt	70	const,test.a.type	13	Using index condition
+drop table t0,t1,t2,t3;

=== added file 'mysql-test/t/join_optimizer.test'
--- a/mysql-test/t/join_optimizer.test	1970-01-01 00:00:00 +0000
+++ b/mysql-test/t/join_optimizer.test	2008-09-12 13:29:34 +0000
@@ -0,0 +1,45 @@
+--disable_warnings
+drop table if exists t0,t1,t2,t3;
+--enable_warnings
+
+--echo #
+--echo # BUG#38049 incorrect rows estimations with references from preceding table
+--echo #
+
+create table t0 (a int);
+insert into t0 values (0),(1),(2),(3),(4),(5),(6),(7),(8),(9);
+
+create table t1 (a varchar(32));
+insert into t1 values ('owner'),('requester'),('admincc'),('cc');
+
+CREATE TABLE t2 (
+  id int(11) NOT NULL,
+  type varchar(32) default NULL,
+  PRIMARY KEY  (id)
+);
+insert into t2 values (1,'owner'), (2,'admincc');
+
+
+CREATE TABLE t3 (
+  id int(11) NOT NULL,
+  domain varchar(32) default NULL,
+  type varchar(32) default NULL,
+  PRIMARY KEY  (id)
+);
+
+set @domain='system';
+set @pk=0;
+INSERT INTO t3 select @pk:=@pk+1, 'system', t1.a from t1;
+INSERT INTO t3 select @pk:=@pk+1, 'queue', t1.a from t1, t0 where t0.a<3;
+INSERT INTO t3 select @pk:=@pk+1, 'ticket', t1.a from t1, t0 A, t0 B, t0 C;
+
+CREATE INDEX groups_d ON t3(domain);
+CREATE INDEX groups_t ON t3(type);
+CREATE INDEX groups_td ON t3(type, domain);
+CREATE INDEX groups_dt ON t3(domain, type);
+--echo For table g this must use ref(groups_dt) and #rows should be around 15 and not 335:
+explain 
+SELECT STRAIGHT_JOIN g.id FROM t2 a, t3 g USE INDEX(groups_dt) 
+WHERE g.domain = 'queue' AND g.type = a.type;
+
+drop table t0,t1,t2,t3;

=== modified file 'sql/sql_select.cc'
--- a/sql/sql_select.cc	2008-08-16 15:22:47 +0000
+++ b/sql/sql_select.cc	2008-09-12 13:29:34 +0000
@@ -5710,7 +5710,8 @@ best_access_path(JOIN      *join,
                 in ReuseRangeEstimateForRef-3.
               */
               if (table->quick_keys.is_set(key) &&
-                  const_part & (1 << table->quick_key_parts[key]) &&
+                  (const_part & ((1 << table->quick_key_parts[key])-1)) ==
+                  ((1 << table->quick_key_parts[key])-1) &&
                   table->quick_n_ranges[key] == 1 &&
                   records > (double) table->quick_rows[key])
               {