[chenjinbao1989]

In add_multi_dqas_hash_agg_path(), the HashAggregates that remove duplicates after the TupleSplit only deduplicate their input; they do not run any aggregate transition or final functions. But DedupCost was filled by get_agg_clause_costs() before they were created, so they were charged the costs of all aggregates in the query. Move the call after them, so the dedup aggregates are costed with empty aggregate costs.

Fixes #ISSUE_Number

What does this PR do?

Type of Change

• Bug fix (non-breaking change)
• New feature (non-breaking change)
• Breaking change (fix or feature with breaking changes)
• Documentation update

Breaking Changes

Test Plan

• Unit tests added/updated
• Integration tests added/updated
• Passed make installcheck
• Passed make -C src/test installcheck-cbdb-parallel

Impact

Performance:

User-facing changes:

Dependencies:

Checklist

• Followed contribution guide
• Added/updated documentation
• Reviewed code for security implications
• This PR contains AI-assisted code generation
• Requested review from cloudberry committers

Additional Context

CI Skip Instructions

---

[my-ship-it]

Thanks for fixing it and clean up FIXMEs, and some comments below:

• The commit message says the get_agg_clause_costs() call is moved, but the diff only deletes it, so please reword it to "remove".

• Please remove the GPDB_14_MERGE_FIXME about DedupCost in gp_dqa.sql/.out (which this PR fixes), and check whether the gp_motion_cost_per_row=2 workaround is still needed.

• No existing test covers this change, so please add a multi-DQA case whose plan changes because of the fix.

• The single-DQA dedup HashAggregates are still costed with ctx->agg_partial_costs /* FIXME */ and have the same overcharge, so please consider fixing them here too.