Commit 2893d11c authored by igor@rurik.mysql.com's avatar igor@rurik.mysql.com
Browse files

Fixed bug #16674. 

The length of the prefix of the pattern string in the LIKE predicate that 
determined the index range to be scanned was calculated incorrectly for
multi-byte character sets. 
As a result of this in 4. 1 the the scanned range was wider then necessary
if the prefix contained not only one-byte characters.  
In 5.0 additionally it caused missing some rows from the result set.
parent 217faf56

mysql-test/r/ctype_utf8.result

+67 −0
Original line number Diff line number Diff line
@@ -1124,3 +1124,70 @@ check table t1; 
Table	Op	Msg_type	Msg_text

test.t1	check	status	OK

drop table t1;

SET NAMES utf8;

CREATE TABLE t1 (

a CHAR(13) DEFAULT '',

INDEX(a)

) ENGINE=MyISAM DEFAULT CHARSET=utf8 COLLATE=utf8_general_ci;

INSERT INTO t1 VALUES 

('Käli Käli 2-4'), ('Käli Käli 2-4'),

('Käli Käli 2+4'), ('Käli Käli 2+4'),

('Käli Käli 2-6'), ('Käli Käli 2-6');

CREATE TABLE t2 (

a CHAR(13) DEFAULT '',

INDEX(a)

) ENGINE=MyISAM DEFAULT CHARSET=latin1 COLLATE=latin1_general_ci;

INSERT INTO t2 VALUES

('Kali Kali 2-4'), ('Kali Kali 2-4'),

('Kali Kali 2+4'), ('Kali Kali 2+4'),

('Kali Kali 2-6'), ('Kali Kali 2-6');

SELECT a FROM t1 WHERE a LIKE 'Käli Käli 2+4';

a

Käli Käli 2+4

Käli Käli 2+4

SELECT a FROM t2 WHERE a LIKE 'Kali Kali 2+4';

a

Kali Kali 2+4

Kali Kali 2+4

EXPLAIN SELECT a FROM t1 WHERE a LIKE 'Käli Käli 2+4';

id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra

1	SIMPLE	t1	range	a	a	40	NULL	2	Using where; Using index

EXPLAIN SELECT a FROM t1 WHERE a = 'Käli Käli 2+4';

id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra

1	SIMPLE	t1	ref	a	a	40	const	2	Using where; Using index

EXPLAIN SELECT a FROM t2 WHERE a LIKE 'Kali Kali 2+4';

id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra

1	SIMPLE	t2	range	a	a	14	NULL	2	Using where; Using index

EXPLAIN SELECT a FROM t2 WHERE a = 'Kali Kali 2+4';

id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra

1	SIMPLE	t2	ref	a	a	14	const	2	Using where; Using index

DROP TABLE t1,t2;

CREATE TABLE t1 (

a char(255) DEFAULT '', 

KEY(a(10))

) ENGINE=MyISAM DEFAULT CHARSET=utf8 COLLATE=utf8_general_ci;

INSERT INTO t1 VALUES ('Käli Käli 2-4');

SELECT * FROM t1 WHERE a LIKE 'Käli Käli 2%';

a

Käli Käli 2-4

INSERT INTO t1 VALUES ('Käli Käli 2-4');

SELECT * FROM t1 WHERE a LIKE 'Käli Käli 2%';

a

Käli Käli 2-4

Käli Käli 2-4

DROP TABLE t1;

CREATE TABLE t1 (

a char(255) DEFAULT ''

) ENGINE=MyISAM DEFAULT CHARSET=utf8 COLLATE=utf8_general_ci;

INSERT INTO t1 VALUES ('Käli Käli 2-4');

INSERT INTO t1 VALUES ('Käli Käli 2-4');

SELECT * FROM t1 WHERE a LIKE 'Käli Käli 2%';

a

Käli Käli 2-4

Käli Käli 2-4

ALTER TABLE t1 ADD KEY (a(10));

SELECT * FROM t1 WHERE a LIKE 'Käli Käli 2%';

a

Käli Käli 2-4

Käli Käli 2-4

DROP TABLE t1;

mysql-test/t/ctype_utf8.test

+55 −0
Original line number Diff line number Diff line
@@ -926,4 +926,59 @@ INSERT INTO t1 VALUES('uUABCDEFGHIGKLMNOPRSTUVWXYZ̈bbbbbbbbbbbbbbbbbbbbbbbbbbbb 
check table t1;

drop table t1;



#

# Bug#16674: LIKE predicate for a utf8 character set column

#



SET NAMES utf8;



CREATE TABLE t1 (

  a CHAR(13) DEFAULT '',

  INDEX(a)

) ENGINE=MyISAM DEFAULT CHARSET=utf8 COLLATE=utf8_general_ci;

INSERT INTO t1 VALUES 

 ('Käli Käli 2-4'), ('Käli Käli 2-4'),

 ('Käli Käli 2+4'), ('Käli Käli 2+4'),

 ('Käli Käli 2-6'), ('Käli Käli 2-6');



CREATE TABLE t2 (

  a CHAR(13) DEFAULT '',

  INDEX(a)

) ENGINE=MyISAM DEFAULT CHARSET=latin1 COLLATE=latin1_general_ci;



INSERT INTO t2 VALUES

 ('Kali Kali 2-4'), ('Kali Kali 2-4'),

 ('Kali Kali 2+4'), ('Kali Kali 2+4'),

 ('Kali Kali 2-6'), ('Kali Kali 2-6');



SELECT a FROM t1 WHERE a LIKE 'Käli Käli 2+4';

SELECT a FROM t2 WHERE a LIKE 'Kali Kali 2+4';



EXPLAIN SELECT a FROM t1 WHERE a LIKE 'Käli Käli 2+4';

EXPLAIN SELECT a FROM t1 WHERE a = 'Käli Käli 2+4';

EXPLAIN SELECT a FROM t2 WHERE a LIKE 'Kali Kali 2+4';

EXPLAIN SELECT a FROM t2 WHERE a = 'Kali Kali 2+4';



DROP TABLE t1,t2;



CREATE TABLE t1 (

  a char(255) DEFAULT '', 

  KEY(a(10))

) ENGINE=MyISAM DEFAULT CHARSET=utf8 COLLATE=utf8_general_ci;

INSERT INTO t1 VALUES ('Käli Käli 2-4');

SELECT * FROM t1 WHERE a LIKE 'Käli Käli 2%';

INSERT INTO t1 VALUES ('Käli Käli 2-4');

SELECT * FROM t1 WHERE a LIKE 'Käli Käli 2%';

DROP TABLE t1;



CREATE TABLE t1 (

  a char(255) DEFAULT ''

) ENGINE=MyISAM DEFAULT CHARSET=utf8 COLLATE=utf8_general_ci;

INSERT INTO t1 VALUES ('Käli Käli 2-4');

INSERT INTO t1 VALUES ('Käli Käli 2-4');

SELECT * FROM t1 WHERE a LIKE 'Käli Käli 2%';

ALTER TABLE t1 ADD KEY (a(10));

SELECT * FROM t1 WHERE a LIKE 'Käli Käli 2%';

DROP TABLE t1;



# End of 4.1 tests

strings/ctype-mb.c

+16 −9
Original line number Diff line number Diff line
@@ -502,21 +502,19 @@ my_bool my_like_range_mb(CHARSET_INFO *cs, 
			 char *min_str,char *max_str,

			 uint *min_length,uint *max_length)

{

  uint mblen;

  const char *end= ptr + ptr_length;

  char *min_org= min_str;

  char *min_end= min_str + res_length;

  char *max_end= max_str + res_length;

  uint charlen= res_length / cs->mbmaxlen;

  uint maxcharlen= res_length / cs->mbmaxlen;



  for (; ptr != end && min_str != min_end && charlen > 0 ; ptr++, charlen--)

  for (; ptr != end && min_str != min_end && maxcharlen ; maxcharlen--)

  {

    /* We assume here that escape, w_any, w_namy are one-byte characters */

    if (*ptr == escape && ptr+1 != end)

    {

      ptr++;                                    /* Skip escape */

      *min_str++= *max_str++ = *ptr;

      continue;

    }

    if (*ptr == w_one || *ptr == w_many)	/* '_' and '%' in SQL */

    else if (*ptr == w_one || *ptr == w_many)   /* '_' and '%' in SQL */

    {

      /* Write min key  */

      *min_length= (uint) (min_str - min_org);
@@ -534,7 +532,16 @@ my_bool my_like_range_mb(CHARSET_INFO *cs, 
      pad_max_char(cs, max_str, max_end);

      return 0;

    }

    *min_str++= *max_str++ = *ptr;

    if ((mblen= my_ismbchar(cs, ptr, end)) > 1)

    {

      if (ptr+mblen > end || min_str+mblen > min_end)

        break;

      while (mblen--)

       *min_str++= *max_str++= *ptr++;

    }

    else

       *min_str++= *max_str++= *ptr++;    



  }

  *min_length= *max_length = (uint) (min_str - min_org);