Skip to content

MSQ: Use frames to back broadcast joinables.#16913

Closed
gianm wants to merge 1 commit intoapache:masterfrom
gianm:msq-broadcast-frame-inline
Closed

MSQ: Use frames to back broadcast joinables.#16913
gianm wants to merge 1 commit intoapache:masterfrom
gianm:msq-broadcast-frame-inline

Conversation

@gianm
Copy link
Contributor

@gianm gianm commented Aug 16, 2024

In BroadcastJoinSegmentMapFnProcessor, use FrameBasedInlineDataSource and FrameBasedIndexedTable to back broadcast joinables, rather than a regular InlineDataSource (which would use Java object arrays).

Reduces memory usage and eliminates a copy while building the joinable.

In BroadcastJoinSegmentMapFnProcessor, use FrameBasedInlineDataSource and
FrameBasedIndexedTable to back broadcast joinables, rather than a regular
InlineDataSource (which would use Java object arrays).

Reduces memory usage and eliminates a copy while building the joinable.
@github-actions github-actions bot added Area - Batch Ingestion Area - Segment Format and Ser/De Area - MSQ For multi stage queries - https://github.com/apache/druid/issues/12262 labels Aug 16, 2024
Copy link
Contributor

@LakshSingla LakshSingla left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I was wondering if there's a better way than converting the incoming frame to columnar frame. #14999 tries to have QueryableIndex for the row-based frames as well that I was using to test the join performance with row based frames. Should we implement QueryableIndex for row based frame so that the conversion isn't required?

@LakshSingla
Copy link
Contributor

AFAIR #14999 worked fine with the frame-based broadcast table and didn't have any performance degradation when compared with using columnar frames.

@gianm
Copy link
Contributor Author

gianm commented Oct 10, 2024

@LakshSingla are you suggesting that in this patch I should use row-based frames rather than columnar?

@github-actions
Copy link

This pull request has been marked as stale due to 60 days of inactivity.
It will be closed in 4 weeks if no further activity occurs. If you think
that's incorrect or this pull request should instead be reviewed, please simply
write any comment. Even if closed, you can still revive the PR at any time or
discuss it on the dev@druid.apache.org list.
Thank you for your contributions.

@github-actions github-actions bot added the stale label Dec 10, 2024
@github-actions
Copy link

github-actions bot commented Jan 8, 2025

This pull request/issue has been closed due to lack of activity. If you think that
is incorrect, or the pull request requires review, you can revive the PR at any time.

@github-actions github-actions bot closed this Jan 8, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Area - Batch Ingestion Area - MSQ For multi stage queries - https://github.com/apache/druid/issues/12262 Area - Segment Format and Ser/De stale

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants