mirror of
https://github.com/ClickHouse/ClickHouse.git
synced 2024-11-16 04:32:33 +00:00
23 lines
1.2 KiB
SQL
23 lines
1.2 KiB
SQL
SET optimize_use_implicit_projections = 0;
|
|
|
|
-- the work for scalar subquery is properly accounted:
|
|
SET max_rows_to_read = 1000000;
|
|
SELECT 1 = (SELECT count() FROM test.hits WHERE NOT ignore(AdvEngineID)); -- { serverError TOO_MANY_ROWS }
|
|
|
|
-- the work for subquery in IN is properly accounted:
|
|
SET max_rows_to_read = 1000000;
|
|
SELECT 1 IN (SELECT count() FROM test.hits WHERE NOT ignore(AdvEngineID)); -- { serverError TOO_MANY_ROWS }
|
|
|
|
-- this query reads from the table twice:
|
|
SET max_rows_to_read = 15000000;
|
|
SELECT count() IN (SELECT count() FROM test.hits WHERE NOT ignore(AdvEngineID)) FROM test.hits WHERE NOT ignore(AdvEngineID); -- { serverError TOO_MANY_ROWS }
|
|
|
|
-- the resources are properly accounted even if the subquery is evaluated in advance to facilitate the index analysis.
|
|
-- this query is using index and filter out the second reading pass.
|
|
SET max_rows_to_read = 1000000;
|
|
SELECT count() FROM test.hits WHERE CounterID > (SELECT count() FROM test.hits WHERE NOT ignore(AdvEngineID)); -- { serverError TOO_MANY_ROWS }
|
|
|
|
-- this query is using index but have to read all the data twice.
|
|
SET max_rows_to_read = 10000000;
|
|
SELECT count() FROM test.hits WHERE CounterID < (SELECT count() FROM test.hits WHERE NOT ignore(AdvEngineID)); -- { serverError TOO_MANY_ROWS }
|