[
https://issues.apache.org/jira/browse/HIVE-16311?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15945390#comment-15945390
]
Matt McCline commented on HIVE-16311:
-------------------------------------
I think TestHiveDecimal is missing a random suite for divide. The philosophy
of TestHiveDecimal is to test against the original HiveDecimal now called
OldHiveDecimal class (which is a thin shell over BigDecimal) as the golden
correctness. So, I assume this change probably needs to match the old results.
Instead of this change, perhaps we need to write a native or fast version of
Decimal division. The current code has an experiment
FastHiveDecimalImpl.fastSingleWordDivision that is a start in that direction...
> Improve the performance for FastHiveDecimalImpl.fastDivide
> ----------------------------------------------------------
>
> Key: HIVE-16311
> URL: https://issues.apache.org/jira/browse/HIVE-16311
> Project: Hive
> Issue Type: Improvement
> Affects Versions: 2.2.0
> Reporter: Colin Ma
> Assignee: Colin Ma
> Fix For: 2.2.0
>
> Attachments: HIVE-16311.001.patch
>
>
> FastHiveDecimalImpl.fastDivide is poor performance when evaluate the
> expression as 12345.67/123.45
> There are 2 points can be improved:
> 1. Don't always use HiveDecimal.MAX_SCALE as scale when do the
> BigDecimal.divide.
> 2. Get the precision for BigInteger in a fast way if possible.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
