hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gunther Hagleitner" <ghagleit...@hortonworks.com>
Subject Re: Review Request 14430: HIVE-5382
Date Wed, 02 Oct 2013 23:09:55 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/14430/#review26638
-----------------------------------------------------------



http://svn.apache.org/repos/asf/hive/trunk/ql/src/test/results/clientpositive/cast_to_int.q.out
<https://reviews.apache.org/r/14430/#comment51885>

    it'd be good to do some testing in non radix 10, right?



http://svn.apache.org/repos/asf/hive/trunk/serde/src/java/org/apache/hadoop/hive/serde2/lazy/LazyInteger.java
<https://reviews.apache.org/r/14430/#comment51883>

    BigDecimal is really slow and I'm not sure why you need this here...



http://svn.apache.org/repos/asf/hive/trunk/serde/src/java/org/apache/hadoop/hive/serde2/lazy/LazyInteger.java
<https://reviews.apache.org/r/14430/#comment51884>

    depending on radix characters don't have to be between '0' - '9'



http://svn.apache.org/repos/asf/hive/trunk/serde/src/java/org/apache/hadoop/hive/serde2/lazy/LazyInteger.java
<https://reviews.apache.org/r/14430/#comment51886>

    i believe e is actually a valid character for some radixes. not sure how that should be
handled... maybe allow exp notation only in radix 10?



http://svn.apache.org/repos/asf/hive/trunk/serde/src/java/org/apache/hadoop/hive/serde2/lazy/LazyInteger.java
<https://reviews.apache.org/r/14430/#comment51887>

    do we have to do this in double? should be able to keep in long. that way we dont' have
to worry about double weirdness (loss of precision, NaN)



http://svn.apache.org/repos/asf/hive/trunk/serde/src/java/org/apache/hadoop/hive/serde2/lazy/LazyLong.java
<https://reviews.apache.org/r/14430/#comment51889>

    i think we're ending up with a lot of duplicated code now between long/int. i think if
we go this route we need to combine these. (I think performance is the only reason why we
have this twice).



http://svn.apache.org/repos/asf/hive/trunk/serde/src/java/org/apache/hadoop/hive/serde2/lazy/LazyLong.java
<https://reviews.apache.org/r/14430/#comment51888>

    not sure why double will be loss less.


- Gunther Hagleitner


On Oct. 2, 2013, 3:39 a.m., Hari Sankar Sivarama Subramaniyan wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/14430/
> -----------------------------------------------------------
> 
> (Updated Oct. 2, 2013, 3:39 a.m.)
> 
> 
> Review request for hive, Ashutosh Chauhan, Gunther Hagleitner, and Thejas Nair.
> 
> 
> Bugs: HIVE-5382
>     https://issues.apache.org/jira/browse/HIVE-5382
> 
> 
> Repository: hive
> 
> 
> Description
> -------
> 
> Allow casting of exponential notation whenever possible. 
> Limitations:
> 1. The exponential part will be limited between MIN_VAL and MAX_VAL of the data types
even if the final result might be within the boundaries.
> 2. Allow loss-less transformation using BigDecimal and Double types.
> 
> 
> Diffs
> -----
> 
>   http://svn.apache.org/repos/asf/hive/trunk/ql/src/test/queries/clientpositive/cast_to_int.q
1528177 
>   http://svn.apache.org/repos/asf/hive/trunk/ql/src/test/results/clientpositive/cast_to_int.q.out
1528177 
>   http://svn.apache.org/repos/asf/hive/trunk/serde/src/java/org/apache/hadoop/hive/serde2/lazy/LazyInteger.java
1528177 
>   http://svn.apache.org/repos/asf/hive/trunk/serde/src/java/org/apache/hadoop/hive/serde2/lazy/LazyLong.java
1528177 
> 
> Diff: https://reviews.apache.org/r/14430/diff/
> 
> 
> Testing
> -------
> 
> Ran basic tests on cast_to_int.
> 
> 
> Thanks,
> 
> Hari Sankar Sivarama Subramaniyan
> 
>


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