[PIG-1770] matches clause problem with chars that have special meaning in dk.brics - #, @ ..

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.8.0
• Fix Version/s: 0.8.1
• Component/s: impl
• Assignee: Thejas Nair
• Reporter: Thejas Nair
• Created: 16/Dec/10 20:10
• Updated: 25/Apr/11 21:27
• Resolved: 11/Mar/11 20:21

Description

When special chars #, @ , and the 'optional' patterns described here - http://www.brics.dk/automaton/doc/dk/brics/automaton/RegExp.html#RegExp%28java.lang.String%29 are used , the regex match fails to work.

This is related to PIG-965.

Example and workaround are as follows -

grunt> cat t.txt                           
asd#asdf
zxcasdf
2#asdf

grunt> l = load 't.txt' as (a : chararray);
grunt> f = filter l by (a matches '.*#.*');
grunt> dump f; 
-- No output, though two rows are expected.

--As a workaround, add a \ to escape the # . This regex is valid even in 0.7 , and it will be even after this bug is fixed (its valid java regex, which has same meaning as above regex).
grunt> f = filter l by (a matches '.*\\#.*');
grunt> dump f; 
asd#asdf
2#asdf

Issue Links

• is related to: PIG-965 PERFORMANCE: optimize common case in matches (PORegex)