[ 
https://issues.apache.org/jira/browse/HIVE-2427?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13144278#comment-13144278
 ] 

jirapos...@reviews.apache.org commented on HIVE-2427:
-----------------------------------------------------


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/1705/
-----------------------------------------------------------

(Updated 2011-11-04 19:48:37.718693)


Review request for hive, Yongqiang He, Ning Zhang, and Siying Dong.


Changes
-------

It's been a while, so I ran svn up and fixed the conflicts.

I ran the tests and verified they passed.


Summary
-------

I added checks before the type conversions for both joins and unions to either 
throw an error or a warning depending if a bigint was going to be converted to 
a double.

I also added a new variable hive.mapred.bigint.comparison.mode, which is set to 
either strict or nonstrict to indicate whether an error or a warning should be 
given when a bigint is converted to a double.  This is instead of the original 
implementation which used the variable hive.mapred.mode.


This addresses bug HIVE-2427.
    https://issues.apache.org/jira/browse/HIVE-2427


Diffs (updated)
-----

  trunk/common/src/java/org/apache/hadoop/hive/conf/HiveConf.java 1197375 
  trunk/conf/hive-default.xml 1197375 
  trunk/ql/src/java/org/apache/hadoop/hive/ql/parse/ErrorMsg.java 1197375 
  trunk/ql/src/java/org/apache/hadoop/hive/ql/parse/SemanticAnalyzer.java 
1197375 
  trunk/ql/src/java/org/apache/hadoop/hive/ql/plan/ExprNodeGenericFuncDesc.java 
1197375 
  trunk/ql/src/test/queries/clientnegative/compare_double_bigint.q 1197375 
  trunk/ql/src/test/queries/clientnegative/compare_double_bigint_join.q 
PRE-CREATION 
  trunk/ql/src/test/queries/clientnegative/compare_double_bigint_union.q 
PRE-CREATION 
  trunk/ql/src/test/queries/clientnegative/compare_string_bigint.q 1197375 
  trunk/ql/src/test/queries/clientnegative/compare_string_bigint_join.q 
PRE-CREATION 
  trunk/ql/src/test/queries/clientnegative/compare_string_bigint_union.q 
PRE-CREATION 
  trunk/ql/src/test/queries/clientpositive/filter_join_breaktask2.q 1197375 
  trunk/ql/src/test/results/clientnegative/compare_double_bigint.q.out 1197375 
  trunk/ql/src/test/results/clientnegative/compare_double_bigint_join.q.out 
PRE-CREATION 
  trunk/ql/src/test/results/clientnegative/compare_double_bigint_union.q.out 
PRE-CREATION 
  trunk/ql/src/test/results/clientnegative/compare_string_bigint.q.out 1197375 
  trunk/ql/src/test/results/clientnegative/compare_string_bigint_join.q.out 
PRE-CREATION 
  trunk/ql/src/test/results/clientnegative/compare_string_bigint_union.q.out 
PRE-CREATION 
  trunk/ql/src/test/results/clientpositive/filter_join_breaktask2.q.out 1197375 

Diff: https://reviews.apache.org/r/1705/diff


Testing
-------

I added several more tests to demonstrate the errors.

I also modified a unit test in which a join was being done which involved a 
comparison between a string and a bigint.  I set 
hive.mapred.bigint.comparison.mode to nonstrict, so the only change was the new 
warning appeared in the output.


Thanks,

Kevin


                
> Warn user that precision is lost when bigint is implicitly cast to double in 
> joins and unions.
> ----------------------------------------------------------------------------------------------
>
>                 Key: HIVE-2427
>                 URL: https://issues.apache.org/jira/browse/HIVE-2427
>             Project: Hive
>          Issue Type: Improvement
>            Reporter: Kevin Wilfong
>            Assignee: Kevin Wilfong
>         Attachments: HIVE-2427.1.patch.txt, HIVE-2427.2.patch.txt, 
> HIVE-2427.3.patch.txt
>
>
> When a bigint is implicitly cast to a double (when a bigint is involved in an 
> equality expression with a string or double as part of a join condition, or 
> when a bigint column is to be combined with a string or double column as part 
> of a union) precision may be lost, resulting in unexpected behavior. Until we 
> fix the underlying issue we should throw an error in strict mode, and a 
> warning in nonstrict mode alerting the user about this.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: 
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Reply via email to