You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit da75266
Browse filesBrowse the repository at this point in the historyBrowse files
minor: move create_group_accumulator out of the legacy aggregate stream (apache#25539)
## Which issue does this PR close?
- Part of apache#25537. Follow-up to apache#22710.
## Rationale for this change
`create_group_accumulator` is defined in `grouped_hash_stream.rs`, the
legacy `GroupedHashAggregateStream` that is only reachable with
`datafusion.execution.enable_migration_aggregate = false` and is
scheduled for removal. It is, however, the one piece of that file the
new code depends on: both `AggregateHashTable`
(`aggregate_hash_table/common.rs`) and `OrderedAggregateTable`
(`aggregate_hash_table/common_ordered.rs`) import it from there.
Moving it next to its real users removes the only dependency of the new
aggregation code on the legacy file, so the legacy stream can later be
deleted without touching the new streams.
## What changes are included in this PR?
Pure code move, no behaviour change.
- Move `create_group_accumulator` to `aggregate_hash_table/common.rs`
and re-export it from `aggregate_hash_table`.
- `grouped_hash_stream.rs` now imports it from there (dependency
direction reversed: legacy → new).
## What is the testing strategy for this PR?
Existing tests; no new behaviour.
## Are there any user-facing changes?
No.
0 commit comments