mirror of
https://github.com/timescale/timescaledb.git
synced 2025-05-28 01:30:29 +08:00
5c263286a3
When inserting into a hypertable using a sub-select clause that involves an aggregate, the insert fails with the error "Aggref found in non-Agg plan node". This happens because the target list from the aggregate sub-select expects an aggregate parent node and we simply reuse the target list when modifying the insert plan with new CustomScan plan nodes. This change creates a modified target list to use with the CustomScan node that avoids this issue.
71 lines
2.8 KiB
SQL
71 lines
2.8 KiB
SQL
\ir include/insert_two_partitions.sql
|
|
|
|
\d+ "_timescaledb_internal".*
|
|
SELECT * FROM _timescaledb_catalog.chunk;
|
|
|
|
SELECT * FROM "two_Partitions" ORDER BY "timeCustom", device_id;
|
|
SELECT * FROM ONLY "two_Partitions";
|
|
|
|
CREATE TABLE error_test(time timestamp, temp float8, device text NOT NULL);
|
|
SELECT create_hypertable('error_test', 'time', 'device', 2);
|
|
|
|
\set QUIET off
|
|
INSERT INTO error_test VALUES ('Mon Mar 20 09:18:20.1 2017', 21.3, 'dev1');
|
|
\set ON_ERROR_STOP 0
|
|
-- generate insert error
|
|
INSERT INTO error_test VALUES ('Mon Mar 20 09:18:22.3 2017', 21.1, NULL);
|
|
\set ON_ERROR_STOP 1
|
|
INSERT INTO error_test VALUES ('Mon Mar 20 09:18:25.7 2017', 22.4, 'dev2');
|
|
\set QUIET on
|
|
SELECT * FROM error_test;
|
|
|
|
--test character(9) partition keys since there were issues with padding causing partitioning errors
|
|
CREATE TABLE tick_character (
|
|
symbol character(9) NOT NULL,
|
|
mid REAL NOT NULL,
|
|
spread REAL NOT NULL,
|
|
time TIMESTAMPTZ NOT NULL
|
|
);
|
|
|
|
SELECT create_hypertable ('tick_character', 'time', 'symbol', 2);
|
|
INSERT INTO tick_character ( symbol, mid, spread, time ) VALUES ( 'GBPJPY', 142.639000, 5.80, 'Mon Mar 20 09:18:22.3 2017') RETURNING time, symbol, mid;
|
|
SELECT * FROM tick_character;
|
|
|
|
|
|
CREATE TABLE many_partitions_test(time timestamp, temp float8, device text NOT NULL);
|
|
SELECT create_hypertable('many_partitions_test', 'time', 'device', 1000);
|
|
--NOTE: how much slower the first two queries are -- they are creating chunks
|
|
INSERT INTO many_partitions_test
|
|
SELECT to_timestamp(ser), ser, ser::text FROM generate_series(1,100) ser;
|
|
INSERT INTO many_partitions_test
|
|
SELECT to_timestamp(ser), ser, ser::text FROM generate_series(101,200) ser;
|
|
INSERT INTO many_partitions_test
|
|
SELECT to_timestamp(ser), ser, (ser-201)::text FROM generate_series(201,300) ser;
|
|
|
|
SELECT * FROM many_partitions_test ORDER BY time DESC LIMIT 2;
|
|
SELECT count(*) FROM many_partitions_test;
|
|
|
|
CREATE TABLE date_col_test(time date, temp float8, device text NOT NULL);
|
|
SELECT create_hypertable('date_col_test', 'time', 'device', 1000, chunk_time_interval => INTERVAL '1 Day');
|
|
INSERT INTO date_col_test
|
|
VALUES ('2001-02-01', 98, 'dev1'),
|
|
('2001-03-02', 98, 'dev1');
|
|
|
|
SELECT * FROM date_col_test WHERE time > '2001-01-01';
|
|
|
|
CREATE TABLE many_partitions_test_1m (time timestamp, temp float8, device text NOT NULL);
|
|
SELECT create_hypertable('many_partitions_test_1m', 'time', 'device', 1000);
|
|
|
|
EXPLAIN (verbose on, costs off)
|
|
INSERT INTO many_partitions_test_1m(time, temp, device)
|
|
SELECT time_bucket('1 minute', time) AS period, avg(temp), device
|
|
FROM many_partitions_test
|
|
GROUP BY period, device;
|
|
|
|
INSERT INTO many_partitions_test_1m(time, temp, device)
|
|
SELECT time_bucket('1 minute', time) AS period, avg(temp), device
|
|
FROM many_partitions_test
|
|
GROUP BY period, device;
|
|
|
|
SELECT * FROM many_partitions_test_1m ORDER BY time, device LIMIT 10;
|