Fix order-preservation in pandas-compat unsorted groupby #16942
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
When sort=False is requested in groupby aggregations and pandas compatibility mode is enabled, we are on the hook to produce the grouped aggregation result in an order which matches the input key order. We previously nearly did this, but the reordering relied on the (incorrect) assumption that when joining two tables with a left join, the resulting gather map for the left table is the identity. This is not the case. To fix this, we must permute the right (result) table gather map by the ordering that makes the left map the identity (AKA, sort by key with the left map as keys) before permuting the result. While here, replace the (bounds-checking) IndexedFrame.take call with usage of the internal (non-bounds-checking) _gather method. This avoids a redundant reduction over the indices, since by construction they are in bounds. - Closes rapidsai#16908
wence-
added
bug
mroeschke
approved these changes
Sep 27, 2024
|
/merge |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
When sort=False is requested in groupby aggregations and pandas compatibility mode is enabled, we are on the hook to produce the grouped aggregation result in an order which matches the input key order. We previously nearly did this, but the reordering relied on the (incorrect) assumption that when joining two tables with a left join, the resulting gather map for the left table is the identity. This is not the case.
To fix this, we must permute the right (result) table gather map by the ordering that makes the left map the identity (AKA, sort by key with the left map as keys) before permuting the result.
While here, replace the (bounds-checking) IndexedFrame.take call with usage of the internal (non-bounds-checking) _gather method. This avoids a redundant reduction over the indices, since by construction they are in bounds.
sort=False#16908Checklist