-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/24404/
-----------------------------------------------------------
(Updated Aug. 7, 2014, 3:40 p.m.)
Review request for hive.
Changes
-------
Fixed the failed test due to the missing update to having.q.out for Tez.
Recreated the diff and uploaded here. Thanks for the review.
Bugs: HIVE-7635
https://issues.apache.org/jira/browse/HIVE-7635
Repository: hive-git
Description
-------
A query having same aggregate functions but in different case (e.g. SELECT key,
COUNT(value) FROM src GROUP BY key HAVING count(value) >= 4) does not work and
throws IndexOutOfBoundsException. The cause is that Hive treats count(value)
and COUNT(value) in this query as two different aggregate expression when
compiling query and generating plan. They are case sensitive.
Diffs (updated)
-----
ql/src/java/org/apache/hadoop/hive/ql/parse/SemanticAnalyzer.java 51838ae
ql/src/test/queries/clientpositive/having.q 5b1aa69
ql/src/test/results/clientpositive/having.q.out d912001
ql/src/test/results/clientpositive/tez/having.q.out e96342d
Diff: https://reviews.apache.org/r/24404/diff/
Testing
-------
1. The fix addressed the failed query with different case in aggregate function
name in the query
2. New unit tests passed
3. patch will be submitted for pre-commit tests
Thanks,
Chaoyu Tang
