[IGNITE-13919] Calcite integration. Single group aggregates

• Type: Improvement
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: 2.15
• Component/s: sql
• Labels: calcite calcite3-required
• Release Note: SQL Calcite: Added optimization for COUNT(column) if column is indexed
• Ignite Flags: Docs Required, Release Notes Required
• Assignee: Aleksey Plekhanov
• Reporter: Taras Ledkov
• Created: 28/Dec/20 13:58
• Updated: 12/Dec/22 14:57
• Resolved: 12/Dec/22 14:57

Description

Investigate the optimization for single group aggregates,
e.g.: SELECT <aggregates> FROM <table>

Now the hash based aggregates are used for this case.
We can obtain a little performance improvement.
Also the special cases (e.g.: SELECT COUNT(<any-columns>) FROM <table>) may be optimized to use table/cache size.

Issue Links

• is part of: IGNITE-12248 Apache Calcite based query execution engine
• is related to: IGNITE-16920 Calcite Engine. COUNT(*) lacks trivial optimization.
• links to: GitHub Pull Request #10390