Bug #73277 REGEXP escape of - not escaped properly in character group
Submitted: 11 Jul 2014 13:23 Modified: 15 Jul 2014 13:55
Reporter: Robert Pazurek Email Updates:
Status: Verified Impact on me:
Category:MySQL Server: Parser Severity:S3 (Non-critical)
Version:5.537-35.1-log, 5.0.97, 5.1.74, 5.5.39, 5.6.21, 5.7.5 OS:Any
Assigned to: CPU Architecture:Any

[11 Jul 2014 13:23] Robert Pazurek
In some cases the - character can not be escaped when in character group, depending on it's position.

How to repeat:
SELECT 'bla-blubb' REGEXP 'bla[2\-1]blubb'
throws: invalid character range (1139)

SELECT 'bla-blubb' REGEXP 'bla[1\-2]blubb'
returns: 0
should return: 1

SELECT 'bla2blubb' REGEXP 'bla[1\-3]blubb'
returns: 1
should return: 0

=> the last two are worse, since they silently behave unexpected!

Workaround: if \- is put to the end of the character group, it is escaped properly:

SELECT 'bla2blubb' REGEXP 'bla[13\-]blubb'
returns: 0 (correct)

SELECT 'bla-blubb' REGEXP 'bla[13\-]blubb'
returns: 1 (correct)

Suggested fix:
Escaping the - with a \ should work anywhere, no matter of the position in a character group.
[15 Jul 2014 13:55] Sveta Smirnova
Thank you for the report.

First of all escape character in regular expression is '\\', not single '\' (http://dev.mysql.com/doc/refman/5.6/en/regexp.html#operator_regexp), but in any case your complain regarding to SELECT 'bla-blubb' REGEXP 'bla[2\\-1]blubb'; is correct and repeatable even if I use '\\'. Bug is verified.