Tracking: query result cache

[RinChanNOWWW]

Summary

Query result cache can help us to get query results faster if the data is not changed for the query.

This is a tracking issue for query result cache.

Basic

• RFC: query result cache #10011
• Write and read result from result cache. feat(query): query result cache #10042
• Support cache of multiple tables query. feat(query): support multi-table query result cache. #10158
• Use AST as the cache key. chore(query): use formatted AST to generate the result cache key. #10165
• Store result cache as parquet. feat: store result cache as parquet file. #10187

Related Tables

• System table: system.query_cache. feat(query): Support query result cache system table and table function #10159
• Table function: SELECT * FROM RESULT_SCAN(LAST_QUERY_ID(-2)); feat(query): query cache table function #10256

Optional

• Garbage collection for result cache.
• Different compression methods for result cache.
• Support result cache for compound SQLs. (INSERT SELECT, COPY FROM, CREATE SELECT)
• Provide a API returns query_id <-> sql list in current session, so that user can quickly get the result corresponding to the previous SQL. cc @flaneur2020

[RinChanNOWWW]

Open new issues if we need the optional features.