The new setting should allow to control query complexity on leaf nodes
excluding the final merging stage on the root-node. For example, distributed
query that reads 1k rows from 5 shards will breach the `max_rows_to_read=5000`,
while effectively every shard reads only 1k rows. With setting `max_rows_to_read_leaf=1500`
this limit won't be reached and query will succeed since every shard reads
not more that ~1k rows.
Previous set of QueryProcessingStage does not allow to do this.
But after WithMergeableStateAfterAggregation had been introduced the
following queries can be optimized too under
optimize_distributed_group_by_sharding_key:
- GROUP BY sharding_key LIMIT
- GROUP BY sharding_key LIMIT BY
- GROUP BY sharding_key ORDER BY
And right now it is still not supports:
- WITH TOTALS (looks like it can be supported)
- WITH ROLLUP (looks like it can be supported)
- WITH CUBE
- SETTINGS extremes=1 (looks like it can be supported)
But will be implemented separatelly.
vX: fixes
v2: fix WITH *
v3: fix extremes
v4: fix LIMIT OFFSET (and make a little bit cleaner)
v5: fix HAVING
v6: fix ORDER BY
v7: rebase against 20.7
v8: move out WithMergeableStateAfterAggregation
v9: add optimize_distributed_group_by_sharding_key into test names
Process query until the stage where the aggregate functions were
calculated and finalized.
It will be used for optimize_distributed_group_by_sharding_key.
v2: fix aliases
v3: Fix protocol ABI breakage due to WithMergeableStateAfterAggregation
Conditions >= for QueryProcessingStage::Enum has been verified, and they
are ok (in InterpreterSelectQuery).