ClickHouse/dbms/tests/integration/test_format_avro_confluent/test.py

86 lines
2.3 KiB
Python
Raw Normal View History

2020-02-03 00:02:19 +00:00
import json
import logging
import io
import pytest
from helpers.cluster import ClickHouseCluster, ClickHouseInstance
import helpers.client
import avro.schema
from confluent.schemaregistry.client import CachedSchemaRegistryClient
from confluent.schemaregistry.serializers import MessageSerializer
logging.getLogger().setLevel(logging.INFO)
logging.getLogger().addHandler(logging.StreamHandler())
@pytest.fixture(scope="module")
def cluster():
try:
cluster = ClickHouseCluster(__file__)
cluster.add_instance("dummy", with_kafka=True)
logging.info("Starting cluster...")
cluster.start()
logging.info("Cluster started")
yield cluster
finally:
cluster.shutdown()
def run_query(instance, query, data=None, settings=None):
2020-02-03 00:02:19 +00:00
# type: (ClickHouseInstance, str, object, dict) -> str
logging.info("Running query '{}'...".format(query))
# use http to force parsing on server
if not data:
data = " " # make POST request
result = instance.http_query(query, data=data, params=settings)
2020-02-03 00:02:19 +00:00
logging.info("Query finished")
return result
def test_select(cluster):
# type: (ClickHouseCluster) -> None
schema_registry_client = cluster.schema_registry_client
serializer = MessageSerializer(schema_registry_client)
schema = avro.schema.make_avsc_object({
'name': 'test_record',
'type': 'record',
'fields': [
{
'name': 'value',
'type': 'long'
}
]
})
buf = io.BytesIO()
for x in range(0, 3):
message = serializer.encode_record_with_schema(
'test_subject', schema, {'value': x}
)
buf.write(message)
data = buf.getvalue()
2020-02-03 00:02:19 +00:00
instance = cluster.instances["dummy"] # type: ClickHouseInstance
schema_registry_url = "http://{}:{}".format(
cluster.schema_registry_host,
cluster.schema_registry_port
)
run_query(instance, "create table avro_data(value Int64) engine = Memory()")
settings = {'format_avro_schema_registry_url': schema_registry_url}
run_query(instance, "insert into avro_data format AvroConfluent", data, settings)
2020-02-03 00:02:19 +00:00
stdout = run_query(instance, "select * from avro_data")
assert list(map(str.split, stdout.splitlines())) == [
["0"],
["1"],
["2"],
]