mirror of
https://github.com/ClickHouse/ClickHouse.git
synced 2024-11-16 12:44:42 +00:00
b75963d370
This PR formats all the `*.py` files found under the `tests/integration` folder. It also reorders the imports and cleans up a bunch of unused imports. The formatting also takes care of other things like wrapping lines and fixing spaces and indents such that the tests look more readable.
98 lines
3.0 KiB
Python
98 lines
3.0 KiB
Python
# pylint: disable=unused-argument
|
|
# pylint: disable=redefined-outer-name
|
|
|
|
import pytest
|
|
from helpers.client import QueryRuntimeException
|
|
from helpers.cluster import ClickHouseCluster
|
|
|
|
cluster = ClickHouseCluster(__file__)
|
|
|
|
node = cluster.add_instance('node', with_zookeeper=True)
|
|
|
|
|
|
@pytest.fixture(scope='module')
|
|
def start_cluster():
|
|
try:
|
|
cluster.start()
|
|
|
|
yield cluster
|
|
finally:
|
|
cluster.shutdown()
|
|
|
|
|
|
def get_counts():
|
|
src = int(node.query("SELECT count() FROM test"))
|
|
a = int(node.query("SELECT count() FROM test_mv_a"))
|
|
b = int(node.query("SELECT count() FROM test_mv_b"))
|
|
c = int(node.query("SELECT count() FROM test_mv_c"))
|
|
return src, a, b, c
|
|
|
|
|
|
def test_basic(start_cluster):
|
|
node.query(
|
|
'''
|
|
CREATE TABLE test (A Int64) ENGINE = ReplicatedMergeTree ('/clickhouse/test/tables/test','1') ORDER BY tuple();
|
|
CREATE MATERIALIZED VIEW test_mv_a Engine=ReplicatedMergeTree ('/clickhouse/test/tables/test_mv_a','1') order by tuple() AS SELECT A FROM test;
|
|
CREATE MATERIALIZED VIEW test_mv_b Engine=ReplicatedMergeTree ('/clickhouse/test/tables/test_mv_b','1') partition by A order by tuple() AS SELECT A FROM test;
|
|
CREATE MATERIALIZED VIEW test_mv_c Engine=ReplicatedMergeTree ('/clickhouse/test/tables/test_mv_c','1') order by tuple() AS SELECT A FROM test;
|
|
INSERT INTO test values(999);
|
|
INSERT INTO test values(999);
|
|
'''
|
|
)
|
|
with pytest.raises(QueryRuntimeException):
|
|
node.query(
|
|
'''
|
|
SET max_partitions_per_insert_block = 3;
|
|
INSERT INTO test SELECT number FROM numbers(10);
|
|
'''
|
|
)
|
|
|
|
old_src, old_a, old_b, old_c = get_counts()
|
|
# number of rows in test_mv_a and test_mv_c depends on order of inserts into views
|
|
assert old_src == 11
|
|
assert old_a in (1, 11)
|
|
assert old_b == 1
|
|
assert old_c in (1, 11)
|
|
|
|
node.query("INSERT INTO test SELECT number FROM numbers(10)")
|
|
src, a, b, c = get_counts()
|
|
# no changes because of deduplication in source table
|
|
assert src == old_src
|
|
assert a == old_a
|
|
assert b == old_b
|
|
assert c == old_c
|
|
|
|
node.query(
|
|
'''
|
|
SET deduplicate_blocks_in_dependent_materialized_views = 1;
|
|
INSERT INTO test SELECT number FROM numbers(10);
|
|
'''
|
|
)
|
|
src, a, b, c = get_counts()
|
|
assert src == 11
|
|
assert a == old_a + 10 # first insert could be succesfull with disabled dedup
|
|
assert b == 11
|
|
assert c == old_c + 10
|
|
|
|
with pytest.raises(QueryRuntimeException):
|
|
node.query(
|
|
'''
|
|
SET max_partitions_per_insert_block = 3;
|
|
SET deduplicate_blocks_in_dependent_materialized_views = 1;
|
|
INSERT INTO test SELECT number FROM numbers(100,10);
|
|
'''
|
|
)
|
|
|
|
node.query(
|
|
'''
|
|
SET deduplicate_blocks_in_dependent_materialized_views = 1;
|
|
INSERT INTO test SELECT number FROM numbers(100,10);
|
|
'''
|
|
)
|
|
|
|
src, a, b, c = get_counts()
|
|
assert src == 21
|
|
assert a == old_a + 20
|
|
assert b == 21
|
|
assert c == old_c + 20
|