Enable Parallel Aggregation for Non-Overlapping Partitioned Data

[gene-bordegaray]

Will finish later

Which issue does this PR close?

• Closes Enable Parallel Aggregation for Non-Overlapping Partitioned Data #18777.

Rationale for this change

What changes are included in this PR?

Are these changes tested?

Benchmark results:

TPCH remains unchange as expected since the files don't use Hive-Partitioning

I created my own benchmark using partitioned files on large datasets and saw these speed-ups:

Query 1: Simple Aggregate by Partition Key

  SELECT f_dkey, COUNT(*), SUM(value), AVG(value)
  FROM hive_facts
  GROUP BY f_dkey

  | Version                   | Time   | Speedup      |
  |---------------------------|--------|--------------|
  | Before (FinalPartitioned) | 0.599s | baseline     |
  | After (SinglePartitioned) | 0.303s | 1.98× faster |

  Query 2: Prefix Matching (Partition + Timestamp)

  SELECT f_dkey, timestamp, COUNT(*), SUM(value)
  FROM hive_facts
  GROUP BY f_dkey, timestamp
  LIMIT 10

  | Version                   | Time   | Speedup      |
  |---------------------------|--------|--------------|
  | Before (FinalPartitioned) | 1.160s | baseline     |
  | After (SinglePartitioned) | 0.852s | 1.36× faster |

Are there any user-facing changes?

[gene-bordegaray]

cc: @NGA-TRAN

[NGA-TRAN]

Thanks @gene-bordegaray — impressive performance improvement! I’ll plan to review this during the first week of December. Adding a feature design description to the PR would be helpful for context.

If you identify any follow‑up work, feel free to open tickets, link this PR there, and continue working on next PRs on top of this branch.