Azat Khuzhin 8703b7863d Fix distributed_group_by_no_merge=2 with GROUP BY ... Before this patch the following query: SELECT assumeNotNull(argMax(dummy, 1)) FROM remote('127.1', system.one) SETTINGS distributed_group_by_no_merge = 2 Leads to: Code: 10. DB::Exception: Received from localhost:9000. DB::Exception: Not found column argMax(dummy, 1) in block: while executing 'INPUT : 0 -> argMax(dummy, 1) UInt8 : 0'. Since it tries to execute function one more time, but shards will not send this column when the query processed with distributed_group_by_no_merge=2 (i.e. up to WithMergeableStateAfterAggregation). v0: no exception v2: execut window functions v3: throw exception, since executing window function in this case will lead to messy output		2021-05-06 00:21:47 +03:00
..
0_stateless	Fix distributed_group_by_no_merge=2 with GROUP BY	2021-05-06 00:21:47 +03:00
1_stateful	Revert "[RFC] Fix memory tracking with min_bytes_to_use_mmap_io"	2021-04-19 01:30:53 +03:00
bugs
__init__.py
.gitignore
conftest.py
query_test.py
server.py
shell_config.sh
skip_list.json	Preserve errors for INSERT into Distributed	2021-04-28 22:33:29 +03:00