-- -- MULTI_ARRAY_AGG -- SET citus.next_shard_id TO 520000; CREATE OR REPLACE FUNCTION array_sort (ANYARRAY) RETURNS ANYARRAY LANGUAGE SQL AS $$ SELECT ARRAY(SELECT unnest($1) ORDER BY 1) $$; -- Check multi_cat_agg() aggregate which is used to implement array_agg() SELECT array_cat_agg(i) FROM (VALUES (ARRAY[1,2]), (NULL), (ARRAY[3,4])) AS t(i); array_cat_agg --------------- {1,2,3,4} (1 row) -- Check that we don't support distinct and order by with array_agg() SELECT array_agg(distinct l_orderkey) FROM lineitem; ERROR: array_agg (distinct) is unsupported SELECT array_agg(l_orderkey ORDER BY l_partkey) FROM lineitem; ERROR: array_agg with order by is unsupported SELECT array_agg(distinct l_orderkey ORDER BY l_orderkey) FROM lineitem; ERROR: array_agg with order by is unsupported -- Check array_agg() for different data types and LIMIT clauses SELECT array_sort(array_agg(l_partkey)) FROM lineitem GROUP BY l_orderkey ORDER BY l_orderkey LIMIT 10; array_sort -------------------------------------------------- {2132,15635,24027,63700,67310,155190} {106170} {4297,19036,29380,62143,128449,183095} {88035} {37531,108570,123927} {139636} {79251,94780,145243,151894,157238,163073,182052} {2743,11615,44161,82704,85811,197921} {33918,60519,61336,137469} {88362,89414,169544} (10 rows) SELECT array_sort(array_agg(l_extendedprice)) FROM lineitem GROUP BY l_orderkey ORDER BY l_orderkey LIMIT 10; array_sort ----------------------------------------------------------------- {13309.60,21168.23,22824.48,28955.64,45983.16,49620.16} {44694.46} {2618.76,28733.64,32986.52,39890.88,46796.47,54058.05} {30690.90} {23678.55,50723.92,73426.50} {61998.31} {6476.15,11594.16,13608.60,31809.96,43058.75,73943.82,81639.88} {2210.32,6582.96,9159.66,47227.60,64605.44,79059.64} {7532.30,40217.23,47344.32,75928.31} {9681.24,17554.68,30875.02} (10 rows) SELECT array_sort(array_agg(l_shipdate)) FROM lineitem GROUP BY l_orderkey ORDER BY l_orderkey LIMIT 10; array_sort -------------------------------------------------------------------------------- {01-29-1996,01-30-1996,03-13-1996,03-30-1996,04-12-1996,04-21-1996} {01-28-1997} {10-29-1993,11-09-1993,12-04-1993,12-14-1993,01-16-1994,02-02-1994} {01-10-1996} {08-08-1994,10-16-1994,10-31-1994} {04-27-1992} {01-15-1996,01-16-1996,02-01-1996,02-10-1996,02-11-1996,03-21-1996,05-07-1996} {07-21-1995,08-04-1995,08-07-1995,08-14-1995,08-28-1995,10-23-1995} {10-29-1993,11-09-1993,12-09-1993,12-09-1993} {10-09-1998,10-23-1998,10-30-1998} (10 rows) SELECT array_sort(array_agg(l_shipmode)) FROM lineitem GROUP BY l_orderkey ORDER BY l_orderkey LIMIT 10; array_sort ---------------------------------------------------------------------------------------------- {"AIR ","FOB ","MAIL ","MAIL ","REG AIR ","TRUCK "} {"RAIL "} {"AIR ","FOB ","RAIL ","RAIL ","SHIP ","TRUCK "} {"REG AIR "} {"AIR ","AIR ","FOB "} {"TRUCK "} {"FOB ","FOB ","FOB ","FOB ","MAIL ","SHIP ","TRUCK "} {"AIR ","AIR ","AIR ","RAIL ","REG AIR ","TRUCK "} {"AIR ","MAIL ","MAIL ","TRUCK "} {"FOB ","FOB ","REG AIR "} (10 rows) -- Check that we can execute array_agg() within other functions SELECT array_length(array_agg(l_orderkey), 1) FROM lineitem; array_length -------------- 12000 (1 row) -- Check that we can execute array_agg() on select queries that hit multiple -- shards and contain different aggregates, filter clauses and other complex -- expressions. Note that the l_orderkey ranges are such that the matching rows -- lie in different shards. SELECT l_quantity, count(*), avg(l_extendedprice), array_sort(array_agg(l_orderkey)) FROM lineitem WHERE l_quantity < 5 AND l_orderkey > 5500 AND l_orderkey < 9500 GROUP BY l_quantity ORDER BY l_quantity; l_quantity | count | avg | array_sort ------------+-------+-----------------------+-------------------------------------------------------------------------------------------------- 1.00 | 17 | 1477.1258823529411765 | {5543,5633,5634,5698,5766,5856,5857,5986,8997,9026,9158,9184,9220,9222,9348,9383,9476} 2.00 | 19 | 3078.4242105263157895 | {5506,5540,5573,5669,5703,5730,5798,5831,5893,5920,5923,9030,9058,9123,9124,9188,9344,9441,9476} 3.00 | 14 | 4714.0392857142857143 | {5509,5543,5605,5606,5827,9124,9157,9184,9223,9254,9349,9414,9475,9477} 4.00 | 19 | 5929.7136842105263158 | {5504,5507,5508,5511,5538,5764,5766,5826,5829,5862,5959,5985,9091,9120,9281,9347,9382,9440,9473} (4 rows) SELECT l_quantity, array_sort(array_agg(extract (month FROM o_orderdate))) AS my_month FROM lineitem, orders WHERE l_orderkey = o_orderkey AND l_quantity < 5 AND l_orderkey > 5500 AND l_orderkey < 9500 GROUP BY l_quantity ORDER BY l_quantity; l_quantity | my_month ------------+------------------------------------------------ 1.00 | {2,3,4,4,4,5,5,5,6,7,7,7,7,9,9,11,11} 2.00 | {1,3,5,5,5,5,6,6,6,7,7,8,10,10,11,11,11,12,12} 3.00 | {3,4,5,6,7,7,8,8,8,9,9,10,11,11} 4.00 | {1,1,1,2,2,2,5,5,6,6,6,6,8,9,10,10,11,11,12} (4 rows) SELECT l_quantity, array_sort(array_agg(l_orderkey * 2 + 1)) FROM lineitem WHERE l_quantity < 5 AND octet_length(l_comment) + octet_length('randomtext'::text) > 40 AND l_orderkey > 5500 AND l_orderkey < 9500 GROUP BY l_quantity ORDER BY l_quantity; l_quantity | array_sort ------------+--------------------------------------------- 1.00 | {11269,11397,11713,11715,11973,18317,18445} 2.00 | {11847,18061,18247,18953} 3.00 | {18249,18315,18699,18951,18955} 4.00 | {11653,11659,18241,18765} (4 rows) -- Check that we can execute array_agg() with an expression containing NULL values SELECT array_agg(case when l_quantity > 20 then l_quantity else NULL end) FROM lineitem WHERE l_orderkey < 10; array_agg -------------------------------------------------------------------------------------------------------------------------------------------------- {NULL,36.00,NULL,28.00,24.00,32.00,38.00,45.00,49.00,27.00,NULL,28.00,26.00,30.00,NULL,26.00,50.00,37.00,NULL,NULL,46.00,28.00,38.00,35.00,NULL} (1 row) -- Check that we return NULL in case there are no input rows to array_agg() SELECT array_agg(l_orderkey) FROM lineitem WHERE l_quantity < 0; array_agg ----------- (1 row)