GitHub user viirya reopened a pull request:
https://github.com/apache/spark/pull/12387
[SPARK-14627][SQL] Avoid shilfting encoder many times in
TypedAggregateExpression
## What changes were proposed in this pull request?
JIRA: https://issues.apache.org/jira/browse/SPARK-14627
Every time we call `TypedAggregateExpression.update` method, we call
encoder's `shift` method. As shift method will copy encoder and underlying
`BoundReference`, we should prepare the shifted encoder in advance, instead of
calling `shift` method every time.
BTW, we can also slightly improve encoder's `shift` method to return itself
when shift delta is zero.
## How was this patch tested?
Existing tests.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/viirya/spark-1 skip-copy-bencoder
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/spark/pull/12387.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #12387
----
commit 05875ba0f65ffec90299bee395a8b3771e353a2d
Author: Liang-Chi Hsieh <[email protected]>
Date: 2016-04-14T06:54:04Z
Skip calling encoder.shift many times in update and do not need to do shift
when delta is zero.
commit faf1ddf87c454414624877fbc56f675fa03bb97f
Author: Liang-Chi Hsieh <[email protected]>
Date: 2016-04-14T08:58:42Z
Revert it since it will be revamped soon.
----
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
