datastore/tests/performance/add_distinct_to_in_clause.xml

<test>
    <settings>
        <allow_experimental_analyzer>1</allow_experimental_analyzer>
        <prefer_localhost_replica>0</prefer_localhost_replica>
        <prefer_global_in_and_join>1</prefer_global_in_and_join>
        <distributed_product_mode>global</distributed_product_mode>
        <max_rows_to_read>1000000000000</max_rows_to_read>
        <read_overflow_mode>break</read_overflow_mode>
        <max_threads>1</max_threads>
    </settings>

    <!-- two distributed tables that point to system.numbers on every shard -->
    <create_query>
        create table local_table_1 (id int) engine = MergeTree order by id
    </create_query>

    <create_query>
        create table local_table_2 (id int) engine = MergeTree order by id
    </create_query>

    <create_query>
        create table distributed_table_1 (id int) engine = Distributed(test_cluster_two_shard_three_replicas_localhost, currentDatabase(), local_table_1)
    </create_query>

    <create_query>
        create table distributed_table_2 (id int) engine = Distributed(test_cluster_two_shard_three_replicas_localhost, currentDatabase(), local_table_2)
    </create_query>

    <fill_query>
        insert into local_table_1 select number from numbers(100000000)
    </fill_query>

    <fill_query>
        insert into local_table_2 select 1 from numbers(100000000)
    </fill_query>

    <fill_query>
        insert into local_table_2 select 2 from numbers(100000000);
    </fill_query>

    <fill_query>
        insert into local_table_2 select 3 from numbers(100000000);
    </fill_query>

    <!-- 1.  optimisation OFF  -->
    <query>
        select id from distributed_table_1 where id in (select id from distributed_table_2) settings enable_add_distinct_to_in_subqueries = 0
    </query>

    <!-- 2.  optimisation ON  -->
    <query>
        select id from distributed_table_1 where id in (select id from distributed_table_2) settings enable_add_distinct_to_in_subqueries = 1
    </query>

    <drop_query>DROP TABLE IF EXISTS distributed_table_1</drop_query>
    <drop_query>DROP TABLE IF EXISTS distributed_table_2</drop_query>
    <drop_query>DROP TABLE IF EXISTS local_table_1</drop_query>
    <drop_query>DROP TABLE IF EXISTS local_table_2</drop_query>
</test>