[
https://issues.apache.org/jira/browse/HIVE-24510?focusedWorklogId=532571&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-532571
]
ASF GitHub Bot logged work on HIVE-24510:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 07/Jan/21 15:59
Start Date: 07/Jan/21 15:59
Worklog Time Spent: 10m
Work Description: abstractdog commented on a change in pull request #1824:
URL: https://github.com/apache/hive/pull/1824#discussion_r553419304
##########
File path:
ql/src/test/results/clientpositive/llap/dynpart_sort_optimization.q.out
##########
@@ -161,20 +168,28 @@ STAGE PLANS:
Map-reduce partition columns: _col0 (type:
string), _col1 (type: tinyint)
Statistics: Num rows: 1 Data size: 24 Basic stats:
COMPLETE Column stats: NONE
value expressions: _col2 (type: smallint), _col3
(type: smallint), _col4 (type: bigint), _col5 (type: bigint), _col6 (type:
binary), _col7 (type: int), _col8 (type: int), _col9 (type: bigint), _col10
(type: binary), _col11 (type: bigint), _col12 (type: bigint), _col13 (type:
bigint), _col14 (type: binary), _col15 (type: float), _col16 (type: float),
_col17 (type: bigint), _col18 (type: binary)
- Reduce Output Operator
- key expressions: _col4 (type: tinyint)
- null sort order: a
- sort order: +
- Map-reduce partition columns: _col4 (type: tinyint)
- Statistics: Num rows: 1 Data size: 24 Basic stats:
COMPLETE Column stats: NONE
- value expressions: _col0 (type: smallint), _col1
(type: int), _col2 (type: bigint), _col3 (type: float)
Execution mode: llap
LLAP IO: all inputs
Reducer 2
+ Execution mode: llap
Review comment:
could you please clarify why the patch has moved this operator from
Reducer3 to Reducer2? basically, I "like" that we eliminated a reducer stage,
however, I don't have an idea whether this is correct not, or what's the root
cause behind it?
as a reference, the history of dynamic partitioning sort optimization (what
I can recall) is: HIVE-6455, HIVE-20703, HIVE-23071
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
Issue Time Tracking
-------------------
Worklog Id: (was: 532571)
Time Spent: 1h 20m (was: 1h 10m)
> Vectorize compute_bit_vector
> ----------------------------
>
> Key: HIVE-24510
> URL: https://issues.apache.org/jira/browse/HIVE-24510
> Project: Hive
> Issue Type: Improvement
> Reporter: Mustafa İman
> Assignee: Mustafa İman
> Priority: Major
> Labels: pull-request-available
> Time Spent: 1h 20m
> Remaining Estimate: 0h
>
> After https://issues.apache.org/jira/browse/HIVE-23530 , almost all compute
> stats functions are vectorizable. Only function that is not vectorizable is
> "compute_bit_vector" for ndv statistics computation. This causes "create
> table as select" and "insert overwrite select" queries to run in
> non-vectorized mode.
> Even a very naive implementation of vectorized compute_bit_vector gives about
> 50% performance improvement on simple "insert overwrite select" queries. That
> is because entire mapper or reducer can run in vectorized mode.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
