Github user gatorsmile commented on the issue:
https://github.com/apache/spark/pull/14322
Yeah, I did not see a noticeable performance difference based on the local
tests I did. Based on the outputs of whole-stage code gen, the number of
instructions is less. Thus, I think it helps a
Github user cloud-fan commented on the issue:
https://github.com/apache/spark/pull/14322
how much benefit can we get by avoiding scan partition columns? Seems that
we just parse the directory string to get the partition values, no IO is needed.
---
If your project is set up for it,
Github user gatorsmile commented on the issue:
https://github.com/apache/spark/pull/14322
cc @marmbrus @cloud-fan @liancheng After history checking, most of codes
are done by you. Thanks!
---
If your project is set up for it, you can reply to this email and have your
reply appear
Github user AmplabJenkins commented on the issue:
https://github.com/apache/spark/pull/14322
Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/62743/
Test PASSed.
---
Github user AmplabJenkins commented on the issue:
https://github.com/apache/spark/pull/14322
Merged build finished. Test PASSed.
---
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
Github user SparkQA commented on the issue:
https://github.com/apache/spark/pull/14322
**[Test build #62743 has
finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/62743/consoleFull)**
for PR 14322 at commit
Github user SparkQA commented on the issue:
https://github.com/apache/spark/pull/14322
**[Test build #62743 has
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/62743/consoleFull)**
for PR 14322 at commit
Github user gatorsmile commented on the issue:
https://github.com/apache/spark/pull/14322
**After the PR changes**, the whole-stage codegen output is like:
```JAVA
== Subtree 1 / 1 ==
*Scan json [value#37L] Format: JSON, InputPaths:
Github user gatorsmile commented on the issue:
https://github.com/apache/spark/pull/14322
**Before the PR changes**, the whole-stage codegen output is like:
```JAVA
== Subtree 1 / 1 ==
*Project [value#37L]
+- *Scan json [value#37L,p1#39,p2#40,p3#41] Format: JSON,