hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "John Sichi" <jsi...@fb.com>
Subject Re: Review Request: HIVE-872: Allow BIGINT constants
Date Tue, 14 Jun 2011 19:22:57 GMT


> On 2011-06-14 19:14:04, John Sichi wrote:
> > Wouldn't it be better to change COALESCE to just check that the two objects have
the same type family (both numeric)?  That's what happens in standard SQL, and then the widest
type is used for the result.  This is what UNION ALL should be doing also, although currently
it uses the same annoying check as COALESCE.
> >

Also, CAST(0 AS BIGINT) already works the same as 0L.


- John


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/889/#review834
-----------------------------------------------------------


On 2011-06-14 19:04:53, Syed Albiz wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/889/
> -----------------------------------------------------------
> 
> (Updated 2011-06-14 19:04:53)
> 
> 
> Review request for hive and John Sichi.
> 
> 
> Summary
> -------
> 
> Added a rule to the lexical grammar to allow BIGINT constants ending with 'L', and a
clause to the TypeCheckProcFactory to ensure it gets interpreted properly.
> 
> 
> This addresses bug HIVE-872.
>     https://issues.apache.org/jira/browse/HIVE-872
> 
> 
> Diffs
> -----
> 
>   ql/src/java/org/apache/hadoop/hive/ql/parse/Hive.g 9161319 
>   ql/src/java/org/apache/hadoop/hive/ql/parse/TypeCheckProcFactory.java dfadb9f 
>   ql/src/test/queries/clientpositive/bigint_const.q PRE-CREATION 
>   ql/src/test/results/clientpositive/bigint_const.q.out PRE-CREATION 
> 
> Diff: https://reviews.apache.org/r/889/diff
> 
> 
> Testing
> -------
> 
> TestCliDriver passes, previous behaviour was to accept bigint constants specified without
'L', which is also preserved, so adding additional tests for this case seems unnecessary.
> 
> 
> Thanks,
> 
> Syed
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message