1-- 2-- CREATE_INDEX 3-- Create ancillary data structures (i.e. indices) 4-- 5 6-- 7-- BTREE 8-- 9CREATE INDEX onek_unique1 ON onek USING btree(unique1 int4_ops); 10 11CREATE INDEX IF NOT EXISTS onek_unique1 ON onek USING btree(unique1 int4_ops); 12 13CREATE INDEX IF NOT EXISTS ON onek USING btree(unique1 int4_ops); 14 15CREATE INDEX onek_unique2 ON onek USING btree(unique2 int4_ops); 16 17CREATE INDEX onek_hundred ON onek USING btree(hundred int4_ops); 18 19CREATE INDEX onek_stringu1 ON onek USING btree(stringu1 name_ops); 20 21CREATE INDEX tenk1_unique1 ON tenk1 USING btree(unique1 int4_ops); 22 23CREATE INDEX tenk1_unique2 ON tenk1 USING btree(unique2 int4_ops); 24 25CREATE INDEX tenk1_hundred ON tenk1 USING btree(hundred int4_ops); 26 27CREATE INDEX tenk1_thous_tenthous ON tenk1 (thousand, tenthous); 28 29CREATE INDEX tenk2_unique1 ON tenk2 USING btree(unique1 int4_ops); 30 31CREATE INDEX tenk2_unique2 ON tenk2 USING btree(unique2 int4_ops); 32 33CREATE INDEX tenk2_hundred ON tenk2 USING btree(hundred int4_ops); 34 35CREATE INDEX rix ON road USING btree (name text_ops); 36 37CREATE INDEX iix ON ihighway USING btree (name text_ops); 38 39CREATE INDEX six ON shighway USING btree (name text_ops); 40 41-- test comments 42COMMENT ON INDEX six_wrong IS 'bad index'; 43COMMENT ON INDEX six IS 'good index'; 44COMMENT ON INDEX six IS NULL; 45 46-- 47-- BTREE ascending/descending cases 48-- 49-- we load int4/text from pure descending data (each key is a new 50-- low key) and name/f8 from pure ascending data (each key is a new 51-- high key). we had a bug where new low keys would sometimes be 52-- "lost". 53-- 54CREATE INDEX bt_i4_index ON bt_i4_heap USING btree (seqno int4_ops); 55 56CREATE INDEX bt_name_index ON bt_name_heap USING btree (seqno name_ops); 57 58CREATE INDEX bt_txt_index ON bt_txt_heap USING btree (seqno text_ops); 59 60CREATE INDEX bt_f8_index ON bt_f8_heap USING btree (seqno float8_ops); 61 62-- 63-- BTREE partial indices 64-- 65CREATE INDEX onek2_u1_prtl ON onek2 USING btree(unique1 int4_ops) 66 where unique1 < 20 or unique1 > 980; 67 68CREATE INDEX onek2_u2_prtl ON onek2 USING btree(unique2 int4_ops) 69 where stringu1 < 'B'; 70 71CREATE INDEX onek2_stu1_prtl ON onek2 USING btree(stringu1 name_ops) 72 where onek2.stringu1 >= 'J' and onek2.stringu1 < 'K'; 73 74-- 75-- GiST (rtree-equivalent opclasses only) 76-- 77CREATE INDEX grect2ind ON fast_emp4000 USING gist (home_base); 78 79CREATE INDEX gpolygonind ON polygon_tbl USING gist (f1); 80 81CREATE INDEX gcircleind ON circle_tbl USING gist (f1); 82 83INSERT INTO POINT_TBL(f1) VALUES (NULL); 84 85CREATE INDEX gpointind ON point_tbl USING gist (f1); 86 87CREATE TEMP TABLE gpolygon_tbl AS 88 SELECT polygon(home_base) AS f1 FROM slow_emp4000; 89INSERT INTO gpolygon_tbl VALUES ( '(1000,0,0,1000)' ); 90INSERT INTO gpolygon_tbl VALUES ( '(0,1000,1000,1000)' ); 91 92CREATE TEMP TABLE gcircle_tbl AS 93 SELECT circle(home_base) AS f1 FROM slow_emp4000; 94 95CREATE INDEX ggpolygonind ON gpolygon_tbl USING gist (f1); 96 97CREATE INDEX ggcircleind ON gcircle_tbl USING gist (f1); 98 99-- 100-- Test GiST indexes 101-- 102 103-- get non-indexed results for comparison purposes 104 105SET enable_seqscan = ON; 106SET enable_indexscan = OFF; 107SET enable_bitmapscan = OFF; 108 109SELECT * FROM fast_emp4000 110 WHERE home_base @ '(200,200),(2000,1000)'::box 111 ORDER BY (home_base[0])[0]; 112 113SELECT count(*) FROM fast_emp4000 WHERE home_base && '(1000,1000,0,0)'::box; 114 115SELECT count(*) FROM fast_emp4000 WHERE home_base IS NULL; 116 117SELECT * FROM polygon_tbl WHERE f1 ~ '((1,1),(2,2),(2,1))'::polygon 118 ORDER BY (poly_center(f1))[0]; 119 120SELECT * FROM circle_tbl WHERE f1 && circle(point(1,-2), 1) 121 ORDER BY area(f1); 122 123SELECT count(*) FROM gpolygon_tbl WHERE f1 && '(1000,1000,0,0)'::polygon; 124 125SELECT count(*) FROM gcircle_tbl WHERE f1 && '<(500,500),500>'::circle; 126 127SELECT count(*) FROM point_tbl WHERE f1 <@ box '(0,0,100,100)'; 128 129SELECT count(*) FROM point_tbl WHERE box '(0,0,100,100)' @> f1; 130 131SELECT count(*) FROM point_tbl WHERE f1 <@ polygon '(0,0),(0,100),(100,100),(50,50),(100,0),(0,0)'; 132 133SELECT count(*) FROM point_tbl WHERE f1 <@ circle '<(50,50),50>'; 134 135SELECT count(*) FROM point_tbl p WHERE p.f1 << '(0.0, 0.0)'; 136 137SELECT count(*) FROM point_tbl p WHERE p.f1 >> '(0.0, 0.0)'; 138 139SELECT count(*) FROM point_tbl p WHERE p.f1 <^ '(0.0, 0.0)'; 140 141SELECT count(*) FROM point_tbl p WHERE p.f1 >^ '(0.0, 0.0)'; 142 143SELECT count(*) FROM point_tbl p WHERE p.f1 ~= '(-5, -12)'; 144 145SELECT * FROM point_tbl ORDER BY f1 <-> '0,1'; 146 147SELECT * FROM point_tbl WHERE f1 IS NULL; 148 149SELECT * FROM point_tbl WHERE f1 IS NOT NULL ORDER BY f1 <-> '0,1'; 150 151SELECT * FROM point_tbl WHERE f1 <@ '(-10,-10),(10,10)':: box ORDER BY f1 <-> '0,1'; 152 153SELECT * FROM gpolygon_tbl ORDER BY f1 <-> '(0,0)'::point LIMIT 10; 154 155SELECT circle_center(f1), round(radius(f1)) as radius FROM gcircle_tbl ORDER BY f1 <-> '(200,300)'::point LIMIT 10; 156 157-- Now check the results from plain indexscan 158SET enable_seqscan = OFF; 159SET enable_indexscan = ON; 160SET enable_bitmapscan = OFF; 161 162EXPLAIN (COSTS OFF) 163SELECT * FROM fast_emp4000 164 WHERE home_base @ '(200,200),(2000,1000)'::box 165 ORDER BY (home_base[0])[0]; 166SELECT * FROM fast_emp4000 167 WHERE home_base @ '(200,200),(2000,1000)'::box 168 ORDER BY (home_base[0])[0]; 169 170EXPLAIN (COSTS OFF) 171SELECT count(*) FROM fast_emp4000 WHERE home_base && '(1000,1000,0,0)'::box; 172SELECT count(*) FROM fast_emp4000 WHERE home_base && '(1000,1000,0,0)'::box; 173 174EXPLAIN (COSTS OFF) 175SELECT count(*) FROM fast_emp4000 WHERE home_base IS NULL; 176SELECT count(*) FROM fast_emp4000 WHERE home_base IS NULL; 177 178EXPLAIN (COSTS OFF) 179SELECT * FROM polygon_tbl WHERE f1 ~ '((1,1),(2,2),(2,1))'::polygon 180 ORDER BY (poly_center(f1))[0]; 181SELECT * FROM polygon_tbl WHERE f1 ~ '((1,1),(2,2),(2,1))'::polygon 182 ORDER BY (poly_center(f1))[0]; 183 184EXPLAIN (COSTS OFF) 185SELECT * FROM circle_tbl WHERE f1 && circle(point(1,-2), 1) 186 ORDER BY area(f1); 187SELECT * FROM circle_tbl WHERE f1 && circle(point(1,-2), 1) 188 ORDER BY area(f1); 189 190EXPLAIN (COSTS OFF) 191SELECT count(*) FROM gpolygon_tbl WHERE f1 && '(1000,1000,0,0)'::polygon; 192SELECT count(*) FROM gpolygon_tbl WHERE f1 && '(1000,1000,0,0)'::polygon; 193 194EXPLAIN (COSTS OFF) 195SELECT count(*) FROM gcircle_tbl WHERE f1 && '<(500,500),500>'::circle; 196SELECT count(*) FROM gcircle_tbl WHERE f1 && '<(500,500),500>'::circle; 197 198EXPLAIN (COSTS OFF) 199SELECT count(*) FROM point_tbl WHERE f1 <@ box '(0,0,100,100)'; 200SELECT count(*) FROM point_tbl WHERE f1 <@ box '(0,0,100,100)'; 201 202EXPLAIN (COSTS OFF) 203SELECT count(*) FROM point_tbl WHERE box '(0,0,100,100)' @> f1; 204SELECT count(*) FROM point_tbl WHERE box '(0,0,100,100)' @> f1; 205 206EXPLAIN (COSTS OFF) 207SELECT count(*) FROM point_tbl WHERE f1 <@ polygon '(0,0),(0,100),(100,100),(50,50),(100,0),(0,0)'; 208SELECT count(*) FROM point_tbl WHERE f1 <@ polygon '(0,0),(0,100),(100,100),(50,50),(100,0),(0,0)'; 209 210EXPLAIN (COSTS OFF) 211SELECT count(*) FROM point_tbl WHERE f1 <@ circle '<(50,50),50>'; 212SELECT count(*) FROM point_tbl WHERE f1 <@ circle '<(50,50),50>'; 213 214EXPLAIN (COSTS OFF) 215SELECT count(*) FROM point_tbl p WHERE p.f1 << '(0.0, 0.0)'; 216SELECT count(*) FROM point_tbl p WHERE p.f1 << '(0.0, 0.0)'; 217 218EXPLAIN (COSTS OFF) 219SELECT count(*) FROM point_tbl p WHERE p.f1 >> '(0.0, 0.0)'; 220SELECT count(*) FROM point_tbl p WHERE p.f1 >> '(0.0, 0.0)'; 221 222EXPLAIN (COSTS OFF) 223SELECT count(*) FROM point_tbl p WHERE p.f1 <^ '(0.0, 0.0)'; 224SELECT count(*) FROM point_tbl p WHERE p.f1 <^ '(0.0, 0.0)'; 225 226EXPLAIN (COSTS OFF) 227SELECT count(*) FROM point_tbl p WHERE p.f1 >^ '(0.0, 0.0)'; 228SELECT count(*) FROM point_tbl p WHERE p.f1 >^ '(0.0, 0.0)'; 229 230EXPLAIN (COSTS OFF) 231SELECT count(*) FROM point_tbl p WHERE p.f1 ~= '(-5, -12)'; 232SELECT count(*) FROM point_tbl p WHERE p.f1 ~= '(-5, -12)'; 233 234EXPLAIN (COSTS OFF) 235SELECT * FROM point_tbl ORDER BY f1 <-> '0,1'; 236SELECT * FROM point_tbl ORDER BY f1 <-> '0,1'; 237 238EXPLAIN (COSTS OFF) 239SELECT * FROM point_tbl WHERE f1 IS NULL; 240SELECT * FROM point_tbl WHERE f1 IS NULL; 241 242EXPLAIN (COSTS OFF) 243SELECT * FROM point_tbl WHERE f1 IS NOT NULL ORDER BY f1 <-> '0,1'; 244SELECT * FROM point_tbl WHERE f1 IS NOT NULL ORDER BY f1 <-> '0,1'; 245 246EXPLAIN (COSTS OFF) 247SELECT * FROM point_tbl WHERE f1 <@ '(-10,-10),(10,10)':: box ORDER BY f1 <-> '0,1'; 248SELECT * FROM point_tbl WHERE f1 <@ '(-10,-10),(10,10)':: box ORDER BY f1 <-> '0,1'; 249 250EXPLAIN (COSTS OFF) 251SELECT * FROM gpolygon_tbl ORDER BY f1 <-> '(0,0)'::point LIMIT 10; 252SELECT * FROM gpolygon_tbl ORDER BY f1 <-> '(0,0)'::point LIMIT 10; 253 254EXPLAIN (COSTS OFF) 255SELECT circle_center(f1), round(radius(f1)) as radius FROM gcircle_tbl ORDER BY f1 <-> '(200,300)'::point LIMIT 10; 256SELECT circle_center(f1), round(radius(f1)) as radius FROM gcircle_tbl ORDER BY f1 <-> '(200,300)'::point LIMIT 10; 257 258-- Now check the results from bitmap indexscan 259SET enable_seqscan = OFF; 260SET enable_indexscan = OFF; 261SET enable_bitmapscan = ON; 262 263EXPLAIN (COSTS OFF) 264SELECT * FROM point_tbl WHERE f1 <@ '(-10,-10),(10,10)':: box ORDER BY f1 <-> '0,1'; 265SELECT * FROM point_tbl WHERE f1 <@ '(-10,-10),(10,10)':: box ORDER BY f1 <-> '0,1'; 266 267RESET enable_seqscan; 268RESET enable_indexscan; 269RESET enable_bitmapscan; 270 271-- 272-- GIN over int[] and text[] 273-- 274-- Note: GIN currently supports only bitmap scans, not plain indexscans 275-- 276 277SET enable_seqscan = OFF; 278SET enable_indexscan = OFF; 279SET enable_bitmapscan = ON; 280 281CREATE INDEX intarrayidx ON array_index_op_test USING gin (i); 282 283explain (costs off) 284SELECT * FROM array_index_op_test WHERE i @> '{32}' ORDER BY seqno; 285 286SELECT * FROM array_index_op_test WHERE i @> '{32}' ORDER BY seqno; 287SELECT * FROM array_index_op_test WHERE i && '{32}' ORDER BY seqno; 288SELECT * FROM array_index_op_test WHERE i @> '{17}' ORDER BY seqno; 289SELECT * FROM array_index_op_test WHERE i && '{17}' ORDER BY seqno; 290SELECT * FROM array_index_op_test WHERE i @> '{32,17}' ORDER BY seqno; 291SELECT * FROM array_index_op_test WHERE i && '{32,17}' ORDER BY seqno; 292SELECT * FROM array_index_op_test WHERE i <@ '{38,34,32,89}' ORDER BY seqno; 293SELECT * FROM array_index_op_test WHERE i = '{47,77}' ORDER BY seqno; 294SELECT * FROM array_index_op_test WHERE i = '{}' ORDER BY seqno; 295SELECT * FROM array_index_op_test WHERE i @> '{}' ORDER BY seqno; 296SELECT * FROM array_index_op_test WHERE i && '{}' ORDER BY seqno; 297SELECT * FROM array_index_op_test WHERE i <@ '{}' ORDER BY seqno; 298SELECT * FROM array_op_test WHERE i = '{NULL}' ORDER BY seqno; 299SELECT * FROM array_op_test WHERE i @> '{NULL}' ORDER BY seqno; 300SELECT * FROM array_op_test WHERE i && '{NULL}' ORDER BY seqno; 301SELECT * FROM array_op_test WHERE i <@ '{NULL}' ORDER BY seqno; 302 303CREATE INDEX textarrayidx ON array_index_op_test USING gin (t); 304 305explain (costs off) 306SELECT * FROM array_index_op_test WHERE t @> '{AAAAAAAA72908}' ORDER BY seqno; 307 308SELECT * FROM array_index_op_test WHERE t @> '{AAAAAAAA72908}' ORDER BY seqno; 309SELECT * FROM array_index_op_test WHERE t && '{AAAAAAAA72908}' ORDER BY seqno; 310SELECT * FROM array_index_op_test WHERE t @> '{AAAAAAAAAA646}' ORDER BY seqno; 311SELECT * FROM array_index_op_test WHERE t && '{AAAAAAAAAA646}' ORDER BY seqno; 312SELECT * FROM array_index_op_test WHERE t @> '{AAAAAAAA72908,AAAAAAAAAA646}' ORDER BY seqno; 313SELECT * FROM array_index_op_test WHERE t && '{AAAAAAAA72908,AAAAAAAAAA646}' ORDER BY seqno; 314SELECT * FROM array_index_op_test WHERE t <@ '{AAAAAAAA72908,AAAAAAAAAAAAAAAAAAA17075,AA88409,AAAAAAAAAAAAAAAAAA36842,AAAAAAA48038,AAAAAAAAAAAAAA10611}' ORDER BY seqno; 315SELECT * FROM array_index_op_test WHERE t = '{AAAAAAAAAA646,A87088}' ORDER BY seqno; 316SELECT * FROM array_index_op_test WHERE t = '{}' ORDER BY seqno; 317SELECT * FROM array_index_op_test WHERE t @> '{}' ORDER BY seqno; 318SELECT * FROM array_index_op_test WHERE t && '{}' ORDER BY seqno; 319SELECT * FROM array_index_op_test WHERE t <@ '{}' ORDER BY seqno; 320 321-- And try it with a multicolumn GIN index 322 323DROP INDEX intarrayidx, textarrayidx; 324 325CREATE INDEX botharrayidx ON array_index_op_test USING gin (i, t); 326 327SELECT * FROM array_index_op_test WHERE i @> '{32}' ORDER BY seqno; 328SELECT * FROM array_index_op_test WHERE i && '{32}' ORDER BY seqno; 329SELECT * FROM array_index_op_test WHERE t @> '{AAAAAAA80240}' ORDER BY seqno; 330SELECT * FROM array_index_op_test WHERE t && '{AAAAAAA80240}' ORDER BY seqno; 331SELECT * FROM array_index_op_test WHERE i @> '{32}' AND t && '{AAAAAAA80240}' ORDER BY seqno; 332SELECT * FROM array_index_op_test WHERE i && '{32}' AND t @> '{AAAAAAA80240}' ORDER BY seqno; 333SELECT * FROM array_index_op_test WHERE t = '{}' ORDER BY seqno; 334SELECT * FROM array_op_test WHERE i = '{NULL}' ORDER BY seqno; 335SELECT * FROM array_op_test WHERE i <@ '{NULL}' ORDER BY seqno; 336 337RESET enable_seqscan; 338RESET enable_indexscan; 339RESET enable_bitmapscan; 340 341-- 342-- Try a GIN index with a lot of items with same key. (GIN creates a posting 343-- tree when there are enough duplicates) 344-- 345CREATE TABLE array_gin_test (a int[]); 346 347INSERT INTO array_gin_test SELECT ARRAY[1, g%5, g] FROM generate_series(1, 10000) g; 348 349CREATE INDEX array_gin_test_idx ON array_gin_test USING gin (a); 350 351SELECT COUNT(*) FROM array_gin_test WHERE a @> '{2}'; 352 353DROP TABLE array_gin_test; 354 355-- 356-- Test GIN index's reloptions 357-- 358CREATE INDEX gin_relopts_test ON array_index_op_test USING gin (i) 359 WITH (FASTUPDATE=on, GIN_PENDING_LIST_LIMIT=128); 360\d+ gin_relopts_test 361 362-- 363-- HASH 364-- 365CREATE INDEX hash_i4_index ON hash_i4_heap USING hash (random int4_ops); 366 367CREATE INDEX hash_name_index ON hash_name_heap USING hash (random name_ops); 368 369CREATE INDEX hash_txt_index ON hash_txt_heap USING hash (random text_ops); 370 371CREATE INDEX hash_f8_index ON hash_f8_heap USING hash (random float8_ops) WITH (fillfactor=60); 372 373CREATE UNLOGGED TABLE unlogged_hash_table (id int4); 374CREATE INDEX unlogged_hash_index ON unlogged_hash_table USING hash (id int4_ops); 375DROP TABLE unlogged_hash_table; 376 377-- CREATE INDEX hash_ovfl_index ON hash_ovfl_heap USING hash (x int4_ops); 378 379-- Test hash index build tuplesorting. Force hash tuplesort using low 380-- maintenance_work_mem setting and fillfactor: 381SET maintenance_work_mem = '1MB'; 382CREATE INDEX hash_tuplesort_idx ON tenk1 USING hash (stringu1 name_ops) WITH (fillfactor = 10); 383EXPLAIN (COSTS OFF) 384SELECT count(*) FROM tenk1 WHERE stringu1 = 'TVAAAA'; 385SELECT count(*) FROM tenk1 WHERE stringu1 = 'TVAAAA'; 386DROP INDEX hash_tuplesort_idx; 387RESET maintenance_work_mem; 388 389 390-- 391-- Test functional index 392-- 393CREATE TABLE func_index_heap (f1 text, f2 text); 394CREATE UNIQUE INDEX func_index_index on func_index_heap (textcat(f1,f2)); 395 396INSERT INTO func_index_heap VALUES('ABC','DEF'); 397INSERT INTO func_index_heap VALUES('AB','CDEFG'); 398INSERT INTO func_index_heap VALUES('QWE','RTY'); 399-- this should fail because of unique index: 400INSERT INTO func_index_heap VALUES('ABCD', 'EF'); 401-- but this shouldn't: 402INSERT INTO func_index_heap VALUES('QWERTY'); 403 404 405-- 406-- Same test, expressional index 407-- 408DROP TABLE func_index_heap; 409CREATE TABLE func_index_heap (f1 text, f2 text); 410CREATE UNIQUE INDEX func_index_index on func_index_heap ((f1 || f2) text_ops); 411 412INSERT INTO func_index_heap VALUES('ABC','DEF'); 413INSERT INTO func_index_heap VALUES('AB','CDEFG'); 414INSERT INTO func_index_heap VALUES('QWE','RTY'); 415-- this should fail because of unique index: 416INSERT INTO func_index_heap VALUES('ABCD', 'EF'); 417-- but this shouldn't: 418INSERT INTO func_index_heap VALUES('QWERTY'); 419 420-- 421-- Test unique index with included columns 422-- 423CREATE TABLE covering_index_heap (f1 int, f2 int, f3 text); 424CREATE UNIQUE INDEX covering_index_index on covering_index_heap (f1,f2) INCLUDE(f3); 425 426INSERT INTO covering_index_heap VALUES(1,1,'AAA'); 427INSERT INTO covering_index_heap VALUES(1,2,'AAA'); 428-- this should fail because of unique index on f1,f2: 429INSERT INTO covering_index_heap VALUES(1,2,'BBB'); 430-- and this shouldn't: 431INSERT INTO covering_index_heap VALUES(1,4,'AAA'); 432-- Try to build index on table that already contains data 433CREATE UNIQUE INDEX covering_pkey on covering_index_heap (f1,f2) INCLUDE(f3); 434-- Try to use existing covering index as primary key 435ALTER TABLE covering_index_heap ADD CONSTRAINT covering_pkey PRIMARY KEY USING INDEX 436covering_pkey; 437DROP TABLE covering_index_heap; 438 439 440-- 441-- Also try building functional, expressional, and partial indexes on 442-- tables that already contain data. 443-- 444create unique index hash_f8_index_1 on hash_f8_heap(abs(random)); 445create unique index hash_f8_index_2 on hash_f8_heap((seqno + 1), random); 446create unique index hash_f8_index_3 on hash_f8_heap(random) where seqno > 1000; 447 448-- 449-- Try some concurrent index builds 450-- 451-- Unfortunately this only tests about half the code paths because there are 452-- no concurrent updates happening to the table at the same time. 453 454CREATE TABLE concur_heap (f1 text, f2 text); 455-- empty table 456CREATE INDEX CONCURRENTLY concur_index1 ON concur_heap(f2,f1); 457CREATE INDEX CONCURRENTLY IF NOT EXISTS concur_index1 ON concur_heap(f2,f1); 458INSERT INTO concur_heap VALUES ('a','b'); 459INSERT INTO concur_heap VALUES ('b','b'); 460-- unique index 461CREATE UNIQUE INDEX CONCURRENTLY concur_index2 ON concur_heap(f1); 462CREATE UNIQUE INDEX CONCURRENTLY IF NOT EXISTS concur_index2 ON concur_heap(f1); 463-- check if constraint is set up properly to be enforced 464INSERT INTO concur_heap VALUES ('b','x'); 465-- check if constraint is enforced properly at build time 466CREATE UNIQUE INDEX CONCURRENTLY concur_index3 ON concur_heap(f2); 467-- test that expression indexes and partial indexes work concurrently 468CREATE INDEX CONCURRENTLY concur_index4 on concur_heap(f2) WHERE f1='a'; 469CREATE INDEX CONCURRENTLY concur_index5 on concur_heap(f2) WHERE f1='x'; 470-- here we also check that you can default the index name 471CREATE INDEX CONCURRENTLY on concur_heap((f2||f1)); 472 473-- You can't do a concurrent index build in a transaction 474BEGIN; 475CREATE INDEX CONCURRENTLY concur_index7 ON concur_heap(f1); 476COMMIT; 477 478-- But you can do a regular index build in a transaction 479BEGIN; 480CREATE INDEX std_index on concur_heap(f2); 481COMMIT; 482 483-- Failed builds are left invalid by VACUUM FULL, fixed by REINDEX 484VACUUM FULL concur_heap; 485REINDEX TABLE concur_heap; 486DELETE FROM concur_heap WHERE f1 = 'b'; 487VACUUM FULL concur_heap; 488\d concur_heap 489REINDEX TABLE concur_heap; 490\d concur_heap 491 492-- 493-- Try some concurrent index drops 494-- 495DROP INDEX CONCURRENTLY "concur_index2"; -- works 496DROP INDEX CONCURRENTLY IF EXISTS "concur_index2"; -- notice 497 498-- failures 499DROP INDEX CONCURRENTLY "concur_index2", "concur_index3"; 500BEGIN; 501DROP INDEX CONCURRENTLY "concur_index5"; 502ROLLBACK; 503 504-- successes 505DROP INDEX CONCURRENTLY IF EXISTS "concur_index3"; 506DROP INDEX CONCURRENTLY "concur_index4"; 507DROP INDEX CONCURRENTLY "concur_index5"; 508DROP INDEX CONCURRENTLY "concur_index1"; 509DROP INDEX CONCURRENTLY "concur_heap_expr_idx"; 510 511\d concur_heap 512 513DROP TABLE concur_heap; 514 515-- 516-- Test ADD CONSTRAINT USING INDEX 517-- 518 519CREATE TABLE cwi_test( a int , b varchar(10), c char); 520 521-- add some data so that all tests have something to work with. 522 523INSERT INTO cwi_test VALUES(1, 2), (3, 4), (5, 6); 524 525CREATE UNIQUE INDEX cwi_uniq_idx ON cwi_test(a , b); 526ALTER TABLE cwi_test ADD primary key USING INDEX cwi_uniq_idx; 527 528\d cwi_test 529\d cwi_uniq_idx 530 531CREATE UNIQUE INDEX cwi_uniq2_idx ON cwi_test(b , a); 532ALTER TABLE cwi_test DROP CONSTRAINT cwi_uniq_idx, 533 ADD CONSTRAINT cwi_replaced_pkey PRIMARY KEY 534 USING INDEX cwi_uniq2_idx; 535 536\d cwi_test 537\d cwi_replaced_pkey 538 539DROP INDEX cwi_replaced_pkey; -- Should fail; a constraint depends on it 540 541DROP TABLE cwi_test; 542 543-- ADD CONSTRAINT USING INDEX is forbidden on partitioned tables 544CREATE TABLE cwi_test(a int) PARTITION BY hash (a); 545create unique index on cwi_test (a); 546alter table cwi_test add primary key using index cwi_test_a_idx ; 547DROP TABLE cwi_test; 548 549-- 550-- Check handling of indexes on system columns 551-- 552CREATE TABLE syscol_table (a INT); 553 554-- System columns cannot be indexed 555CREATE INDEX ON syscolcol_table (ctid); 556 557-- nor used in expressions 558CREATE INDEX ON syscol_table ((ctid >= '(1000,0)')); 559 560-- nor used in predicates 561CREATE INDEX ON syscol_table (a) WHERE ctid >= '(1000,0)'; 562 563DROP TABLE syscol_table; 564 565-- 566-- Tests for IS NULL/IS NOT NULL with b-tree indexes 567-- 568 569SELECT unique1, unique2 INTO onek_with_null FROM onek; 570INSERT INTO onek_with_null (unique1,unique2) VALUES (NULL, -1), (NULL, NULL); 571CREATE UNIQUE INDEX onek_nulltest ON onek_with_null (unique2,unique1); 572 573SET enable_seqscan = OFF; 574SET enable_indexscan = ON; 575SET enable_bitmapscan = ON; 576 577SELECT count(*) FROM onek_with_null WHERE unique1 IS NULL; 578SELECT count(*) FROM onek_with_null WHERE unique1 IS NULL AND unique2 IS NULL; 579SELECT count(*) FROM onek_with_null WHERE unique1 IS NOT NULL; 580SELECT count(*) FROM onek_with_null WHERE unique1 IS NULL AND unique2 IS NOT NULL; 581SELECT count(*) FROM onek_with_null WHERE unique1 IS NOT NULL AND unique1 > 500; 582SELECT count(*) FROM onek_with_null WHERE unique1 IS NULL AND unique1 > 500; 583 584DROP INDEX onek_nulltest; 585 586CREATE UNIQUE INDEX onek_nulltest ON onek_with_null (unique2 desc,unique1); 587 588SELECT count(*) FROM onek_with_null WHERE unique1 IS NULL; 589SELECT count(*) FROM onek_with_null WHERE unique1 IS NULL AND unique2 IS NULL; 590SELECT count(*) FROM onek_with_null WHERE unique1 IS NOT NULL; 591SELECT count(*) FROM onek_with_null WHERE unique1 IS NULL AND unique2 IS NOT NULL; 592SELECT count(*) FROM onek_with_null WHERE unique1 IS NOT NULL AND unique1 > 500; 593SELECT count(*) FROM onek_with_null WHERE unique1 IS NULL AND unique1 > 500; 594 595DROP INDEX onek_nulltest; 596 597CREATE UNIQUE INDEX onek_nulltest ON onek_with_null (unique2 desc nulls last,unique1); 598 599SELECT count(*) FROM onek_with_null WHERE unique1 IS NULL; 600SELECT count(*) FROM onek_with_null WHERE unique1 IS NULL AND unique2 IS NULL; 601SELECT count(*) FROM onek_with_null WHERE unique1 IS NOT NULL; 602SELECT count(*) FROM onek_with_null WHERE unique1 IS NULL AND unique2 IS NOT NULL; 603SELECT count(*) FROM onek_with_null WHERE unique1 IS NOT NULL AND unique1 > 500; 604SELECT count(*) FROM onek_with_null WHERE unique1 IS NULL AND unique1 > 500; 605 606DROP INDEX onek_nulltest; 607 608CREATE UNIQUE INDEX onek_nulltest ON onek_with_null (unique2 nulls first,unique1); 609 610SELECT count(*) FROM onek_with_null WHERE unique1 IS NULL; 611SELECT count(*) FROM onek_with_null WHERE unique1 IS NULL AND unique2 IS NULL; 612SELECT count(*) FROM onek_with_null WHERE unique1 IS NOT NULL; 613SELECT count(*) FROM onek_with_null WHERE unique1 IS NULL AND unique2 IS NOT NULL; 614SELECT count(*) FROM onek_with_null WHERE unique1 IS NOT NULL AND unique1 > 500; 615SELECT count(*) FROM onek_with_null WHERE unique1 IS NULL AND unique1 > 500; 616 617DROP INDEX onek_nulltest; 618 619-- Check initial-positioning logic too 620 621CREATE UNIQUE INDEX onek_nulltest ON onek_with_null (unique2); 622 623SET enable_seqscan = OFF; 624SET enable_indexscan = ON; 625SET enable_bitmapscan = OFF; 626 627SELECT unique1, unique2 FROM onek_with_null 628 ORDER BY unique2 LIMIT 2; 629SELECT unique1, unique2 FROM onek_with_null WHERE unique2 >= -1 630 ORDER BY unique2 LIMIT 2; 631SELECT unique1, unique2 FROM onek_with_null WHERE unique2 >= 0 632 ORDER BY unique2 LIMIT 2; 633 634SELECT unique1, unique2 FROM onek_with_null 635 ORDER BY unique2 DESC LIMIT 2; 636SELECT unique1, unique2 FROM onek_with_null WHERE unique2 >= -1 637 ORDER BY unique2 DESC LIMIT 2; 638SELECT unique1, unique2 FROM onek_with_null WHERE unique2 < 999 639 ORDER BY unique2 DESC LIMIT 2; 640 641RESET enable_seqscan; 642RESET enable_indexscan; 643RESET enable_bitmapscan; 644 645DROP TABLE onek_with_null; 646 647-- 648-- Check bitmap index path planning 649-- 650 651EXPLAIN (COSTS OFF) 652SELECT * FROM tenk1 653 WHERE thousand = 42 AND (tenthous = 1 OR tenthous = 3 OR tenthous = 42); 654SELECT * FROM tenk1 655 WHERE thousand = 42 AND (tenthous = 1 OR tenthous = 3 OR tenthous = 42); 656 657EXPLAIN (COSTS OFF) 658SELECT count(*) FROM tenk1 659 WHERE hundred = 42 AND (thousand = 42 OR thousand = 99); 660SELECT count(*) FROM tenk1 661 WHERE hundred = 42 AND (thousand = 42 OR thousand = 99); 662 663-- 664-- Check behavior with duplicate index column contents 665-- 666 667CREATE TABLE dupindexcols AS 668 SELECT unique1 as id, stringu2::text as f1 FROM tenk1; 669CREATE INDEX dupindexcols_i ON dupindexcols (f1, id, f1 text_pattern_ops); 670ANALYZE dupindexcols; 671 672EXPLAIN (COSTS OFF) 673 SELECT count(*) FROM dupindexcols 674 WHERE f1 BETWEEN 'WA' AND 'ZZZ' and id < 1000 and f1 ~<~ 'YX'; 675SELECT count(*) FROM dupindexcols 676 WHERE f1 BETWEEN 'WA' AND 'ZZZ' and id < 1000 and f1 ~<~ 'YX'; 677 678-- 679-- Check ordering of =ANY indexqual results (bug in 9.2.0) 680-- 681 682vacuum tenk1; -- ensure we get consistent plans here 683 684explain (costs off) 685SELECT unique1 FROM tenk1 686WHERE unique1 IN (1,42,7) 687ORDER BY unique1; 688 689SELECT unique1 FROM tenk1 690WHERE unique1 IN (1,42,7) 691ORDER BY unique1; 692 693explain (costs off) 694SELECT thousand, tenthous FROM tenk1 695WHERE thousand < 2 AND tenthous IN (1001,3000) 696ORDER BY thousand; 697 698SELECT thousand, tenthous FROM tenk1 699WHERE thousand < 2 AND tenthous IN (1001,3000) 700ORDER BY thousand; 701 702SET enable_indexonlyscan = OFF; 703 704explain (costs off) 705SELECT thousand, tenthous FROM tenk1 706WHERE thousand < 2 AND tenthous IN (1001,3000) 707ORDER BY thousand; 708 709SELECT thousand, tenthous FROM tenk1 710WHERE thousand < 2 AND tenthous IN (1001,3000) 711ORDER BY thousand; 712 713RESET enable_indexonlyscan; 714 715-- 716-- Check elimination of constant-NULL subexpressions 717-- 718 719explain (costs off) 720 select * from tenk1 where (thousand, tenthous) in ((1,1001), (null,null)); 721 722-- 723-- Check matching of boolean index columns to WHERE conditions and sort keys 724-- 725 726create temp table boolindex (b bool, i int, unique(b, i), junk float); 727 728explain (costs off) 729 select * from boolindex order by b, i limit 10; 730explain (costs off) 731 select * from boolindex where b order by i limit 10; 732explain (costs off) 733 select * from boolindex where b = true order by i desc limit 10; 734explain (costs off) 735 select * from boolindex where not b order by i limit 10; 736explain (costs off) 737 select * from boolindex where b is true order by i desc limit 10; 738explain (costs off) 739 select * from boolindex where b is false order by i desc limit 10; 740 741-- 742-- REINDEX (VERBOSE) 743-- 744CREATE TABLE reindex_verbose(id integer primary key); 745\set VERBOSITY terse \\ -- suppress machine-dependent details 746REINDEX (VERBOSE) TABLE reindex_verbose; 747\set VERBOSITY default 748DROP TABLE reindex_verbose; 749 750-- 751-- REINDEX CONCURRENTLY 752-- 753CREATE TABLE concur_reindex_tab (c1 int); 754-- REINDEX 755REINDEX TABLE concur_reindex_tab; -- notice 756REINDEX TABLE CONCURRENTLY concur_reindex_tab; -- notice 757ALTER TABLE concur_reindex_tab ADD COLUMN c2 text; -- add toast index 758-- Normal index with integer column 759CREATE UNIQUE INDEX concur_reindex_ind1 ON concur_reindex_tab(c1); 760-- Normal index with text column 761CREATE INDEX concur_reindex_ind2 ON concur_reindex_tab(c2); 762-- UNIQUE index with expression 763CREATE UNIQUE INDEX concur_reindex_ind3 ON concur_reindex_tab(abs(c1)); 764-- Duplicate column names 765CREATE INDEX concur_reindex_ind4 ON concur_reindex_tab(c1, c1, c2); 766-- Create table for check on foreign key dependence switch with indexes swapped 767ALTER TABLE concur_reindex_tab ADD PRIMARY KEY USING INDEX concur_reindex_ind1; 768CREATE TABLE concur_reindex_tab2 (c1 int REFERENCES concur_reindex_tab); 769INSERT INTO concur_reindex_tab VALUES (1, 'a'); 770INSERT INTO concur_reindex_tab VALUES (2, 'a'); 771-- Reindex concurrently of exclusion constraint currently not supported 772CREATE TABLE concur_reindex_tab3 (c1 int, c2 int4range, EXCLUDE USING gist (c2 WITH &&)); 773INSERT INTO concur_reindex_tab3 VALUES (3, '[1,2]'); 774REINDEX INDEX CONCURRENTLY concur_reindex_tab3_c2_excl; -- error 775REINDEX TABLE CONCURRENTLY concur_reindex_tab3; -- succeeds with warning 776INSERT INTO concur_reindex_tab3 VALUES (4, '[2,4]'); 777-- Check materialized views 778CREATE MATERIALIZED VIEW concur_reindex_matview AS SELECT * FROM concur_reindex_tab; 779REINDEX INDEX CONCURRENTLY concur_reindex_ind1; 780REINDEX TABLE CONCURRENTLY concur_reindex_tab; 781REINDEX TABLE CONCURRENTLY concur_reindex_matview; 782-- Check that comments are preserved 783CREATE TABLE testcomment (i int); 784CREATE INDEX testcomment_idx1 ON testcomment (i); 785COMMENT ON INDEX testcomment_idx1 IS 'test comment'; 786SELECT obj_description('testcomment_idx1'::regclass, 'pg_class'); 787REINDEX TABLE testcomment; 788SELECT obj_description('testcomment_idx1'::regclass, 'pg_class'); 789REINDEX TABLE CONCURRENTLY testcomment ; 790SELECT obj_description('testcomment_idx1'::regclass, 'pg_class'); 791DROP TABLE testcomment; 792-- Partitions 793-- Create some partitioned tables 794CREATE TABLE concur_reindex_part (c1 int, c2 int) PARTITION BY RANGE (c1); 795CREATE TABLE concur_reindex_part_0 PARTITION OF concur_reindex_part 796 FOR VALUES FROM (0) TO (10) PARTITION BY list (c2); 797CREATE TABLE concur_reindex_part_0_1 PARTITION OF concur_reindex_part_0 798 FOR VALUES IN (1); 799CREATE TABLE concur_reindex_part_0_2 PARTITION OF concur_reindex_part_0 800 FOR VALUES IN (2); 801-- This partitioned table will have no partitions. 802CREATE TABLE concur_reindex_part_10 PARTITION OF concur_reindex_part 803 FOR VALUES FROM (10) TO (20) PARTITION BY list (c2); 804-- Create some partitioned indexes 805CREATE INDEX concur_reindex_part_index ON ONLY concur_reindex_part (c1); 806CREATE INDEX concur_reindex_part_index_0 ON ONLY concur_reindex_part_0 (c1); 807ALTER INDEX concur_reindex_part_index ATTACH PARTITION concur_reindex_part_index_0; 808-- This partitioned index will have no partitions. 809CREATE INDEX concur_reindex_part_index_10 ON ONLY concur_reindex_part_10 (c1); 810ALTER INDEX concur_reindex_part_index ATTACH PARTITION concur_reindex_part_index_10; 811CREATE INDEX concur_reindex_part_index_0_1 ON ONLY concur_reindex_part_0_1 (c1); 812ALTER INDEX concur_reindex_part_index_0 ATTACH PARTITION concur_reindex_part_index_0_1; 813CREATE INDEX concur_reindex_part_index_0_2 ON ONLY concur_reindex_part_0_2 (c1); 814ALTER INDEX concur_reindex_part_index_0 ATTACH PARTITION concur_reindex_part_index_0_2; 815SELECT relid, parentrelid, level FROM pg_partition_tree('concur_reindex_part_index') 816 ORDER BY relid, level; 817-- REINDEX fails for partitioned indexes 818REINDEX INDEX concur_reindex_part_index_10; 819REINDEX INDEX CONCURRENTLY concur_reindex_part_index_10; 820-- REINDEX is a no-op for partitioned tables 821REINDEX TABLE concur_reindex_part_10; 822REINDEX TABLE CONCURRENTLY concur_reindex_part_10; 823SELECT relid, parentrelid, level FROM pg_partition_tree('concur_reindex_part_index') 824 ORDER BY relid, level; 825-- REINDEX should preserve dependencies of partition tree. 826REINDEX INDEX CONCURRENTLY concur_reindex_part_index_0_1; 827REINDEX INDEX CONCURRENTLY concur_reindex_part_index_0_2; 828SELECT relid, parentrelid, level FROM pg_partition_tree('concur_reindex_part_index') 829 ORDER BY relid, level; 830REINDEX TABLE CONCURRENTLY concur_reindex_part_0_1; 831REINDEX TABLE CONCURRENTLY concur_reindex_part_0_2; 832SELECT relid, parentrelid, level FROM pg_partition_tree('concur_reindex_part_index') 833 ORDER BY relid, level; 834DROP TABLE concur_reindex_part; 835 836-- Check errors 837-- Cannot run inside a transaction block 838BEGIN; 839REINDEX TABLE CONCURRENTLY concur_reindex_tab; 840COMMIT; 841REINDEX TABLE CONCURRENTLY pg_database; -- no shared relation 842REINDEX TABLE CONCURRENTLY pg_class; -- no catalog relations 843REINDEX SYSTEM CONCURRENTLY postgres; -- not allowed for SYSTEM 844-- Warns about catalog relations 845REINDEX SCHEMA CONCURRENTLY pg_catalog; 846 847-- Check the relation status, there should not be invalid indexes 848\d concur_reindex_tab 849DROP MATERIALIZED VIEW concur_reindex_matview; 850DROP TABLE concur_reindex_tab, concur_reindex_tab2, concur_reindex_tab3; 851 852-- Check handling of invalid indexes 853CREATE TABLE concur_reindex_tab4 (c1 int); 854INSERT INTO concur_reindex_tab4 VALUES (1), (1), (2); 855-- This trick creates an invalid index. 856CREATE UNIQUE INDEX CONCURRENTLY concur_reindex_ind5 ON concur_reindex_tab4 (c1); 857-- Reindexing concurrently this index fails with the same failure. 858-- The extra index created is itself invalid, and can be dropped. 859REINDEX INDEX CONCURRENTLY concur_reindex_ind5; 860\d concur_reindex_tab4 861DROP INDEX concur_reindex_ind5_ccnew; 862-- This makes the previous failure go away, so the index can become valid. 863DELETE FROM concur_reindex_tab4 WHERE c1 = 1; 864-- The invalid index is not processed when running REINDEX TABLE. 865REINDEX TABLE CONCURRENTLY concur_reindex_tab4; 866\d concur_reindex_tab4 867-- But it is fixed with REINDEX INDEX. 868REINDEX INDEX CONCURRENTLY concur_reindex_ind5; 869\d concur_reindex_tab4 870DROP TABLE concur_reindex_tab4; 871 872-- 873-- REINDEX SCHEMA 874-- 875REINDEX SCHEMA schema_to_reindex; -- failure, schema does not exist 876CREATE SCHEMA schema_to_reindex; 877SET search_path = 'schema_to_reindex'; 878CREATE TABLE table1(col1 SERIAL PRIMARY KEY); 879INSERT INTO table1 SELECT generate_series(1,400); 880CREATE TABLE table2(col1 SERIAL PRIMARY KEY, col2 TEXT NOT NULL); 881INSERT INTO table2 SELECT generate_series(1,400), 'abc'; 882CREATE INDEX ON table2(col2); 883CREATE MATERIALIZED VIEW matview AS SELECT col1 FROM table2; 884CREATE INDEX ON matview(col1); 885CREATE VIEW view AS SELECT col2 FROM table2; 886CREATE TABLE reindex_before AS 887SELECT oid, relname, relfilenode, relkind, reltoastrelid 888 FROM pg_class 889 where relnamespace = (SELECT oid FROM pg_namespace WHERE nspname = 'schema_to_reindex'); 890INSERT INTO reindex_before 891SELECT oid, 'pg_toast_TABLE', relfilenode, relkind, reltoastrelid 892FROM pg_class WHERE oid IN 893 (SELECT reltoastrelid FROM reindex_before WHERE reltoastrelid > 0); 894INSERT INTO reindex_before 895SELECT oid, 'pg_toast_TABLE_index', relfilenode, relkind, reltoastrelid 896FROM pg_class where oid in 897 (select indexrelid from pg_index where indrelid in 898 (select reltoastrelid from reindex_before where reltoastrelid > 0)); 899REINDEX SCHEMA schema_to_reindex; 900CREATE TABLE reindex_after AS SELECT oid, relname, relfilenode, relkind 901 FROM pg_class 902 where relnamespace = (SELECT oid FROM pg_namespace WHERE nspname = 'schema_to_reindex'); 903SELECT b.relname, 904 b.relkind, 905 CASE WHEN a.relfilenode = b.relfilenode THEN 'relfilenode is unchanged' 906 ELSE 'relfilenode has changed' END 907 FROM reindex_before b JOIN pg_class a ON b.oid = a.oid 908 ORDER BY 1; 909REINDEX SCHEMA schema_to_reindex; 910BEGIN; 911REINDEX SCHEMA schema_to_reindex; -- failure, cannot run in a transaction 912END; 913 914-- concurrently 915REINDEX SCHEMA CONCURRENTLY schema_to_reindex; 916 917-- Failure for unauthorized user 918CREATE ROLE regress_reindexuser NOLOGIN; 919SET SESSION ROLE regress_reindexuser; 920REINDEX SCHEMA schema_to_reindex; 921 922-- Clean up 923RESET ROLE; 924DROP ROLE regress_reindexuser; 925DROP SCHEMA schema_to_reindex CASCADE; 926