GitHub user yaooqinn opened a pull request:
https://github.com/apache/spark/pull/15071
[WIP][SPARK-17517][SQL]Improve generated Code for BroadcastHashJoinExec
## What changes were proposed in this pull request?
For current `BroadcastHashJoinExec`, we generate join code for key is not
unique like this:
```java
while (matches.hasnext)
matched = matches.next
check and read stream side row fields
check and read build side row fieldes
reset result row
write stream side row fields to result row
write stream side row fields to result row
```
For some cases, we don't need to check/read/write the steam side repeatedly
in such while circle, e.g. `Inner Join with BuildRight`, or `BuildLeft && all
left side fields are fixed length` and so on. we may generate the code as below:
```java
check and read stream side row fields
reset result row
write stream side row fields to result row
while (matches.hasnext)
matched = matches.next
check and read build side row fieldes
write stream side row fields to result row
```
## How was this patch tested?
todo
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/yaooqinn/spark bhj-codegen
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/spark/pull/15071.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 #15071
----
commit bb731f2bc318ddad03e6543bf45cd5ff7e775206
Author: Kent Yao <[email protected]>
Date: 2016-09-13T03:14:46Z
init commit
----
---
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]
