Apply BroadcastMode key projections before interpreting key expressions in subqueries [databricks] #6268
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Outdated suggestions cannot be applied.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.
Fixes #6232. When the GPU broadcasts for a hash join, it does not broadcast a hash table like the CPU does, rather it just broadcasts the original build table. This is normally fine, but it can cause problems in DPP/subquery usage where the planning assumes that the HashedRelation key expressions have already been applied to the broadcasted data. In the GPU case this has not happened, and thus those expressions can be critical to proper interpretation of the broadcast data. #6232 shows a query where the HashedRelation keys are rearranging the original data, and failure to evaluate those expressions results in the wrong column being used in the subquery exec.
To fix it, we grab any expressions found on the BroadcastMode and evaluate them in the GpuSubqueryExec before trying to apply the normal BoundReference projection to extract the desired key.