mirror of https://github.com/citusdata/citus.git
244 lines
8.1 KiB
Plaintext
244 lines
8.1 KiB
Plaintext
--
|
|
-- MULTI_AGG_DISTINCT
|
|
--
|
|
SET citus.coordinator_aggregation_strategy TO 'disabled';
|
|
-- Create a new range partitioned lineitem table and load data into it
|
|
CREATE TABLE lineitem_range (
|
|
l_orderkey bigint not null,
|
|
l_partkey integer not null,
|
|
l_suppkey integer not null,
|
|
l_linenumber integer not null,
|
|
l_quantity decimal(15, 2) not null,
|
|
l_extendedprice decimal(15, 2) not null,
|
|
l_discount decimal(15, 2) not null,
|
|
l_tax decimal(15, 2) not null,
|
|
l_returnflag char(1) not null,
|
|
l_linestatus char(1) not null,
|
|
l_shipdate date not null,
|
|
l_commitdate date not null,
|
|
l_receiptdate date not null,
|
|
l_shipinstruct char(25) not null,
|
|
l_shipmode char(10) not null,
|
|
l_comment varchar(44) not null );
|
|
SELECT create_distributed_table('lineitem_range', 'l_orderkey', 'range');
|
|
create_distributed_table
|
|
---------------------------------------------------------------------
|
|
|
|
(1 row)
|
|
|
|
SELECT master_create_empty_shard('lineitem_range') AS new_shard_id
|
|
\gset
|
|
UPDATE pg_dist_shard SET shardminvalue = 1, shardmaxvalue = 5986
|
|
WHERE shardid = :new_shard_id;
|
|
SELECT master_create_empty_shard('lineitem_range') AS new_shard_id
|
|
\gset
|
|
UPDATE pg_dist_shard SET shardminvalue = 8997, shardmaxvalue = 14947
|
|
WHERE shardid = :new_shard_id;
|
|
\set lineitem_1_data_file :abs_srcdir '/data/lineitem.1.data'
|
|
\set lineitem_2_data_file :abs_srcdir '/data/lineitem.2.data'
|
|
COPY lineitem_range FROM :'lineitem_1_data_file' with delimiter '|';
|
|
COPY lineitem_range FROM :'lineitem_2_data_file' with delimiter '|';
|
|
-- Run aggregate(distinct) on partition column for range partitioned table
|
|
SELECT count(distinct l_orderkey) FROM lineitem_range;
|
|
count
|
|
---------------------------------------------------------------------
|
|
2985
|
|
(1 row)
|
|
|
|
SELECT avg(distinct l_orderkey) FROM lineitem_range;
|
|
avg
|
|
---------------------------------------------------------------------
|
|
7463.9474036850921273
|
|
(1 row)
|
|
|
|
-- Run count(distinct) on join between a range partitioned table and a single
|
|
-- sharded table. For this test, we also change a config setting to ensure that
|
|
-- we don't repartition any of the tables during the query.
|
|
SELECT p_partkey, count(distinct l_orderkey) FROM lineitem_range, part
|
|
WHERE l_partkey = p_partkey
|
|
GROUP BY p_partkey
|
|
ORDER BY p_partkey LIMIT 10;
|
|
p_partkey | count
|
|
---------------------------------------------------------------------
|
|
18 | 1
|
|
79 | 1
|
|
91 | 1
|
|
149 | 2
|
|
175 | 1
|
|
179 | 1
|
|
182 | 1
|
|
195 | 1
|
|
204 | 1
|
|
222 | 1
|
|
(10 rows)
|
|
|
|
-- Check that we support more complex expressions.
|
|
SELECT count(distinct (l_orderkey)) FROM lineitem_range;
|
|
count
|
|
---------------------------------------------------------------------
|
|
2985
|
|
(1 row)
|
|
|
|
SELECT count(distinct (l_orderkey + 1)) FROM lineitem_range;
|
|
count
|
|
---------------------------------------------------------------------
|
|
2985
|
|
(1 row)
|
|
|
|
SELECT count(distinct (l_orderkey % 5)) FROM lineitem_range;
|
|
count
|
|
---------------------------------------------------------------------
|
|
5
|
|
(1 row)
|
|
|
|
-- count(distinct) on non-partition column is allowed
|
|
SELECT count(distinct l_partkey) FROM lineitem_range;
|
|
count
|
|
---------------------------------------------------------------------
|
|
11661
|
|
(1 row)
|
|
|
|
SELECT count(distinct (l_partkey + 1)) FROM lineitem_range;
|
|
count
|
|
---------------------------------------------------------------------
|
|
11661
|
|
(1 row)
|
|
|
|
SELECT count(distinct (l_partkey % 5)) FROM lineitem_range;
|
|
count
|
|
---------------------------------------------------------------------
|
|
5
|
|
(1 row)
|
|
|
|
-- Now test append partitioned tables. First run count(distinct) on a single
|
|
-- sharded table.
|
|
SELECT count(distinct p_mfgr) FROM part;
|
|
count
|
|
---------------------------------------------------------------------
|
|
5
|
|
(1 row)
|
|
|
|
SELECT p_mfgr, count(distinct p_partkey) FROM part GROUP BY p_mfgr ORDER BY p_mfgr;
|
|
p_mfgr | count
|
|
---------------------------------------------------------------------
|
|
Manufacturer#1 | 193
|
|
Manufacturer#2 | 190
|
|
Manufacturer#3 | 228
|
|
Manufacturer#4 | 204
|
|
Manufacturer#5 | 185
|
|
(5 rows)
|
|
|
|
-- We support count(distinct) queries on append partitioned tables
|
|
-- both on partition column, and non-partition column.
|
|
SELECT count(distinct o_orderkey), count(distinct o_custkey) FROM orders;
|
|
count | count
|
|
---------------------------------------------------------------------
|
|
2985 | 923
|
|
(1 row)
|
|
|
|
-- Hash partitioned tables:
|
|
CREATE TABLE lineitem_hash (
|
|
l_orderkey bigint not null,
|
|
l_partkey integer not null,
|
|
l_suppkey integer not null,
|
|
l_linenumber integer not null,
|
|
l_quantity decimal(15, 2) not null,
|
|
l_extendedprice decimal(15, 2) not null,
|
|
l_discount decimal(15, 2) not null,
|
|
l_tax decimal(15, 2) not null,
|
|
l_returnflag char(1) not null,
|
|
l_linestatus char(1) not null,
|
|
l_shipdate date not null,
|
|
l_commitdate date not null,
|
|
l_receiptdate date not null,
|
|
l_shipinstruct char(25) not null,
|
|
l_shipmode char(10) not null,
|
|
l_comment varchar(44) not null );
|
|
SET citus.shard_replication_factor TO 1;
|
|
SELECT create_distributed_table('lineitem_hash', 'l_orderkey', 'hash');
|
|
create_distributed_table
|
|
---------------------------------------------------------------------
|
|
|
|
(1 row)
|
|
|
|
COPY lineitem_hash FROM :'lineitem_1_data_file' with delimiter '|';
|
|
COPY lineitem_hash FROM :'lineitem_2_data_file' with delimiter '|';
|
|
-- aggregate(distinct) on partition column is allowed
|
|
SELECT count(distinct l_orderkey) FROM lineitem_hash;
|
|
count
|
|
---------------------------------------------------------------------
|
|
2985
|
|
(1 row)
|
|
|
|
SELECT avg(distinct l_orderkey) FROM lineitem_hash;
|
|
avg
|
|
---------------------------------------------------------------------
|
|
7463.9474036850921273
|
|
(1 row)
|
|
|
|
-- Check that we support more complex expressions.
|
|
SELECT count(distinct (l_orderkey)) FROM lineitem_hash;
|
|
count
|
|
---------------------------------------------------------------------
|
|
2985
|
|
(1 row)
|
|
|
|
SELECT count(distinct (l_orderkey + 1)) FROM lineitem_hash;
|
|
count
|
|
---------------------------------------------------------------------
|
|
2985
|
|
(1 row)
|
|
|
|
SELECT count(distinct (l_orderkey % 5)) FROM lineitem_hash;
|
|
count
|
|
---------------------------------------------------------------------
|
|
5
|
|
(1 row)
|
|
|
|
-- count(distinct) on non-partition column is allowed
|
|
SELECT count(distinct l_partkey) FROM lineitem_hash;
|
|
count
|
|
---------------------------------------------------------------------
|
|
11661
|
|
(1 row)
|
|
|
|
SELECT count(distinct (l_partkey + 1)) FROM lineitem_hash;
|
|
count
|
|
---------------------------------------------------------------------
|
|
11661
|
|
(1 row)
|
|
|
|
SELECT count(distinct (l_partkey % 5)) FROM lineitem_hash;
|
|
count
|
|
---------------------------------------------------------------------
|
|
5
|
|
(1 row)
|
|
|
|
-- agg(distinct) is allowed if we group by partition column
|
|
SELECT l_orderkey, count(distinct l_partkey) INTO hash_results FROM lineitem_hash GROUP BY l_orderkey;
|
|
SELECT l_orderkey, count(distinct l_partkey) INTO range_results FROM lineitem_range GROUP BY l_orderkey;
|
|
-- they should return the same results
|
|
SELECT * FROM hash_results h, range_results r WHERE h.l_orderkey = r.l_orderkey AND h.count != r.count;
|
|
l_orderkey | count | l_orderkey | count
|
|
---------------------------------------------------------------------
|
|
(0 rows)
|
|
|
|
-- count(distinct) is allowed if we group by non-partition column
|
|
SELECT l_partkey, count(distinct l_orderkey) INTO hash_results_np FROM lineitem_hash GROUP BY l_partkey;
|
|
SELECT l_partkey, count(distinct l_orderkey) INTO range_results_np FROM lineitem_range GROUP BY l_partkey;
|
|
-- they should return the same results
|
|
SELECT * FROM hash_results_np h, range_results_np r WHERE h.l_partkey = r.l_partkey AND h.count != r.count;
|
|
l_partkey | count | l_partkey | count
|
|
---------------------------------------------------------------------
|
|
(0 rows)
|
|
|
|
-- other agg(distinct) are not allowed on non-partition columns even they are grouped
|
|
-- on non-partition columns
|
|
SELECT SUM(distinct l_partkey) FROM lineitem_hash;
|
|
ERROR: cannot compute aggregate (distinct)
|
|
DETAIL: table partitioning is unsuitable for aggregate (distinct)
|
|
SELECT l_shipmode, sum(distinct l_partkey) FROM lineitem_hash GROUP BY l_shipmode;
|
|
ERROR: cannot compute aggregate (distinct)
|
|
DETAIL: table partitioning is unsuitable for aggregate (distinct)
|
|
DROP TABLE lineitem_hash;
|