Bug #52080 Make REGEXP to work properly with a multibyte character sets

• Submitted: 16 Mar 2010 6:57
• Modified: 27 Aug 2010 9:59
• Reporter: Pavel Sirovatsky
• Status: Duplicate
• Category: MySQL Server: General
• Severity: S4 (Feature request)
• Version: 5.0.27
• OS: Any
• CPU Architecture: Any
• Tags: REGEXP

[16 Mar 2010 6:57] Pavel Sirovatsky

Description:
REGEXP gives wrong result when try to collate non-english (cyrillic) charachters. utf8_unicode_ci is using for collation.

How to repeat:
SET NAMES utf8;

SELECT t.name
FROM (SELECT 'Я' as name) as t
WHERE t.name COLLATE utf8_unicode_ci LIKE 'я';

/*This query works fine*/

SELECT t.name
FROM (SELECT 'Я' as name) as t
WHERE t.name COLLATE utf8_unicode_ci REGEXP 'я';

/*This query don't gives a result*/

[16 Mar 2010 8:32] Valeriy Kravchuk

While this is easily repeatable, our manual (http://dev.mysql.com/doc/refman/5.0/en/regexp.html#operator_regexp) explains this IMHO:

"Warning

The REGEXP and RLIKE operators work in byte-wise fashion, so they are not multi-byte safe and may produce unexpected results with multi-byte character sets. In addition, these operators compare characters by their byte values and accented characters may not compare as equal even if a given collation treats them as equal."

[17 Mar 2010 7:38] Pavel Sirovatsky

change status to feature request

[17 Mar 2010 7:45] Valeriy Kravchuk

Making REGEXP to work properly with a multibyte character sets sounds like a reasonable and nice feature request.

[27 Aug 2010 9:59] Valeriy Kravchuk

Duplicate of Bug#30241.