[PIG-1935] New logical plan: Should not push up filter in front of Bincond

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.8.0
• Fix Version/s: 0.8.1
• Component/s: impl
• Hadoop Flags: Reviewed
• Assignee: Daniel Dai
• Reporter: Daniel Dai
• Created: 25/Mar/11 01:33
• Updated: 25/Apr/11 21:27
• Resolved: 04/Apr/11 20:46

Description

The following script produce wrong result:

data = LOAD 'data.txt' using PigStorage() as (referrer:chararray, canonical_url:chararray, ip:chararray);
best_url = FOREACH data GENERATE ((canonical_url != '' and canonical_url is not null) ? canonical_url : referrer) AS url, ip;
filtered = FILTER best_url BY url == 'badsite.com';
dump filtered;

data.txt:
badsite.com 127.0.0.1
goodsite.com/1?foo=true goodsite.com 127.0.0.1

Expected:
(badsite.com,127.0.0.1)

We get nothing.

Thanks Corbin Hoenes for reporting.