-
Notifications
You must be signed in to change notification settings - Fork 28.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[SPARK-29544][CORE] collect the runtime statistics of row count in map stage #26309
Conversation
@cloud-fan This is the first PR of optimizing skewed partition. Please help me review. Thanks. |
Test build #112889 has finished for PR 26309 at commit
|
Test build #112893 has finished for PR 26309 at commit
|
Test build #112904 has finished for PR 26309 at commit
|
Test build #112980 has finished for PR 26309 at commit
|
retest this please |
Test build #112999 has finished for PR 26309 at commit
|
@cloud-fan please help retest again. The error may be not related. Thanks. |
@cloud-fan please help retest again. Thanks. |
retest this please |
Test build #113086 has finished for PR 26309 at commit
|
Retest this please |
BTW, @JkSelf .
|
Test build #113736 has finished for PR 26309 at commit
|
We're closing this PR because it hasn't been updated in a while. This isn't a judgement on the merit of the PR in any way. It's just a way of keeping the PR queue manageable. |
What changes were proposed in this pull request?
Similar with the approach of collecting data size, this PR collect the row count info when shuffle write and wrap it in MapStatus, then driver can get the row count info from the returned MapStatus.
Why are the changes needed?
In order to optimize the skewed partition, we need collect the row count statistics in map stage.
Does this PR introduce any user-facing change?
No
How was this patch tested?
unit tests