`array_agg` is very slow

[neilconway]

Describe the bug

Running a simple benchmark (5000 groups, 5000 int64 values per group), array_agg in DataFusion is roughly 30x slower than DuckDB and 60x slower than ClickHouse.

To Reproduce

No response

Expected behavior

No response

Additional context

No response

[neilconway]

take

[alamb]

See also

• Implement GroupsAccumulator for array_agg aggregation function #10145

[alamb]

• I think this is a dupe of Slow aggregrate query with array_agg, Polars is 4 times faster for equal query #17446
  Let's continue the discussion there

[neilconway]

@alamb Thanks, Andrew! I missed that this issue was previously reported and worked on. At least my approach isn't radically different than the prior consensus 😅

[alamb]

Even better is that since this keeps coming up it is a signal of how important it is