[
https://issues.apache.org/jira/browse/HIVE-5771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13876529#comment-13876529
]
Ted Xu commented on HIVE-5771:
------------------------------
Hi [[email protected]],
Your points are valid, thanks! Here is my thinking of those issues:
* smb_mapjoin_18.q & smb_mapjoin_25.q: those problems are introduced by
constant propagate optimizer (CPO) conflicting with *Bucketing Sorting
ReduceSink Optimizer (BSRO)*. I tried apply BSRO before CPO and the issue seems
fixed.
* groupby_sort_1.q & groupby_sort_skew_1.q: those are because of CPO
conflicting with *Groupby Optimizer (GO)*, apply it before CPO also fixes
issue. In fact I'm wondering if it is safe to reorder those optimizers, making
it GO->BSRO->CPO.
* decimal.q & pcr.q: I disabled these two cases because of some issue I still
not figured out. My local machine told me to patch a piece of output data like
'0.0040' to '0,004', but it is still '0.0040' in hudson server. I guess it is
an environment issue.
I will update the patch as soon as I validated the above fixes.
> Constant propagation optimizer for Hive
> ---------------------------------------
>
> Key: HIVE-5771
> URL: https://issues.apache.org/jira/browse/HIVE-5771
> Project: Hive
> Issue Type: Improvement
> Components: Query Processor
> Reporter: Ted Xu
> Assignee: Ted Xu
> Attachments: HIVE-5771.1.patch, HIVE-5771.2.patch, HIVE-5771.3.patch,
> HIVE-5771.4.patch, HIVE-5771.5.patch, HIVE-5771.6.patch, HIVE-5771.patch
>
>
> Currently there is no constant folding/propagation optimizer, all expressions
> are evaluated at runtime.
> HIVE-2470 did a great job on evaluating constants on UDF initializing phase,
> however, it is still a runtime evaluation and it doesn't propagate constants
> from a subquery to outside.
> It may reduce I/O and accelerate process if we introduce such an optimizer.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
