背景

《PostgreSQL 设计优化case - 大宽表任意字段组合查询索引如何选择(btree, gin, rum) - (含单个索引列数超过32列的方法)》

《PostgreSQL 任意字段数组合 AND\OR 条件,指定返回结果条数,构造测试数据算法举例》

《PostgreSQL ADHoc(任意字段组合)查询(rums索引加速) - 非字典化,普通、数组等组合字段生成新数组》

《PostgreSQL 实践 - 实时广告位推荐 2 (任意字段组合、任意维度组合搜索、输出TOP-K)》

《PostgreSQL 实践 - 实时广告位推荐 1 (任意字段组合、任意维度组合搜索、输出TOP-K)》

《PostgreSQL ADHoc(任意字段组合)查询 与 字典化 (rum索引加速) - 实践与方案1》

《PostgreSQL 如何高效解决 按任意字段分词检索的问题 - case 1》

《HTAP数据库 PostgreSQL 场景与性能测试之 20 - (OLAP) 用户画像圈人场景 - 多个字段任意组合条件筛选与透视》

《PostgreSQL 多字段任意组合搜索的性能》

1亿记录,128个字段,任意字段组合查询。性能如何?

PG凭什么可以搞定大数据量的任意字段组合实时搜索?

《PostgreSQL 并行计算解说 汇总》

《PostgreSQL 9种索引的原理和应用场景》

例子

1、测试表

  1. do language plpgsql $$
  2. declare
  3. sql text;
  4. begin
  5. sql := 'create unlogged table test(id serial primary key,';
  6. for i in 1..64 loop
  7. sql := sql||' c'||i||' int default random()*100,';
  8. end loop;
  9. for i in 65..128 loop
  10. sql := sql||' c'||i||' int default random()*1000000,';
  11. end loop;
  12. sql := rtrim(sql,',');
  13. sql := sql||')';
  14. execute sql;
  15. end;
  16. $$;

2、写入1亿数据

  1. vi test.sql
  2. insert into test (c1) select random()*100 from generate_series(1,100);
  3. nohup pgbench -M prepared -n -r -P 1 -f ./test.sql -c 50 -j 50 -t 20000 >/dev/null 2>&1 &

3、写完后的大小

  1. postgres=# \dt+ test
  2. List of relations
  3. Schema | Name | Type | Owner | Size | Description
  4. --------+------+-------+----------+-------+-------------
  5. public | test | table | postgres | 55 GB |
  6. (1 row)
  7. postgres=# select count(*) from test;
  8. count
  9. -----------
  10. 100000000
  11. (1 row)

4、高效率创建索引

  1. vi idx.sql
  2. vacuum (analyze,verbose) test;
  3. set maintenance_work_mem='8GB';
  4. set max_parallel_workers=128;
  5. set max_parallel_workers_per_gather=32;
  6. set min_parallel_index_scan_size=0;
  7. set min_parallel_table_scan_size=0;
  8. set parallel_setup_cost=0;
  9. set parallel_tuple_cost=0;
  10. set max_parallel_maintenance_workers=16;
  11. alter table test set (parallel_workers=64);
  12. do language plpgsql $$
  13. declare
  14. sql text;
  15. begin
  16. for i in 1..128 loop
  17. execute format('create index idx_test_%s on test (c%s) %s', i, i, 'tablespace tbs_8001');
  18. end loop;
  19. end;
  20. $$;
  21. vacuum (analyze,verbose) test;
  22. nohup psql -f ./idx.sql >/dev/null 2>&1 &

5、建完索引后

  1. postgres=# \d+ test
  2. Unlogged table "public.test"
  3. Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
  4. --------+---------+-----------+----------+------------------------------------------+---------+--------------+-------------
  5. id | integer | | not null | nextval('test_id_seq'::regclass) | plain | |
  6. c1 | integer | | | (random() * (100)::double precision) | plain | |
  7. c2 | integer | | | (random() * (100)::double precision) | plain | |
  8. c3 | integer | | | (random() * (100)::double precision) | plain | |
  9. c4 | integer | | | (random() * (100)::double precision) | plain | |
  10. c5 | integer | | | (random() * (100)::double precision) | plain | |
  11. c6 | integer | | | (random() * (100)::double precision) | plain | |
  12. c7 | integer | | | (random() * (100)::double precision) | plain | |
  13. c8 | integer | | | (random() * (100)::double precision) | plain | |
  14. c9 | integer | | | (random() * (100)::double precision) | plain | |
  15. c10 | integer | | | (random() * (100)::double precision) | plain | |
  16. c11 | integer | | | (random() * (100)::double precision) | plain | |
  17. c12 | integer | | | (random() * (100)::double precision) | plain | |
  18. c13 | integer | | | (random() * (100)::double precision) | plain | |
  19. c14 | integer | | | (random() * (100)::double precision) | plain | |
  20. c15 | integer | | | (random() * (100)::double precision) | plain | |
  21. c16 | integer | | | (random() * (100)::double precision) | plain | |
  22. c17 | integer | | | (random() * (100)::double precision) | plain | |
  23. c18 | integer | | | (random() * (100)::double precision) | plain | |
  24. c19 | integer | | | (random() * (100)::double precision) | plain | |
  25. c20 | integer | | | (random() * (100)::double precision) | plain | |
  26. c21 | integer | | | (random() * (100)::double precision) | plain | |
  27. c22 | integer | | | (random() * (100)::double precision) | plain | |
  28. c23 | integer | | | (random() * (100)::double precision) | plain | |
  29. c24 | integer | | | (random() * (100)::double precision) | plain | |
  30. c25 | integer | | | (random() * (100)::double precision) | plain | |
  31. c26 | integer | | | (random() * (100)::double precision) | plain | |
  32. c27 | integer | | | (random() * (100)::double precision) | plain | |
  33. c28 | integer | | | (random() * (100)::double precision) | plain | |
  34. c29 | integer | | | (random() * (100)::double precision) | plain | |
  35. c30 | integer | | | (random() * (100)::double precision) | plain | |
  36. c31 | integer | | | (random() * (100)::double precision) | plain | |
  37. c32 | integer | | | (random() * (100)::double precision) | plain | |
  38. c33 | integer | | | (random() * (100)::double precision) | plain | |
  39. c34 | integer | | | (random() * (100)::double precision) | plain | |
  40. c35 | integer | | | (random() * (100)::double precision) | plain | |
  41. c36 | integer | | | (random() * (100)::double precision) | plain | |
  42. c37 | integer | | | (random() * (100)::double precision) | plain | |
  43. c38 | integer | | | (random() * (100)::double precision) | plain | |
  44. c39 | integer | | | (random() * (100)::double precision) | plain | |
  45. c40 | integer | | | (random() * (100)::double precision) | plain | |
  46. c41 | integer | | | (random() * (100)::double precision) | plain | |
  47. c42 | integer | | | (random() * (100)::double precision) | plain | |
  48. c43 | integer | | | (random() * (100)::double precision) | plain | |
  49. c44 | integer | | | (random() * (100)::double precision) | plain | |
  50. c45 | integer | | | (random() * (100)::double precision) | plain | |
  51. c46 | integer | | | (random() * (100)::double precision) | plain | |
  52. c47 | integer | | | (random() * (100)::double precision) | plain | |
  53. c48 | integer | | | (random() * (100)::double precision) | plain | |
  54. c49 | integer | | | (random() * (100)::double precision) | plain | |
  55. c50 | integer | | | (random() * (100)::double precision) | plain | |
  56. c51 | integer | | | (random() * (100)::double precision) | plain | |
  57. c52 | integer | | | (random() * (100)::double precision) | plain | |
  58. c53 | integer | | | (random() * (100)::double precision) | plain | |
  59. c54 | integer | | | (random() * (100)::double precision) | plain | |
  60. c55 | integer | | | (random() * (100)::double precision) | plain | |
  61. c56 | integer | | | (random() * (100)::double precision) | plain | |
  62. c57 | integer | | | (random() * (100)::double precision) | plain | |
  63. c58 | integer | | | (random() * (100)::double precision) | plain | |
  64. c59 | integer | | | (random() * (100)::double precision) | plain | |
  65. c60 | integer | | | (random() * (100)::double precision) | plain | |
  66. c61 | integer | | | (random() * (100)::double precision) | plain | |
  67. c62 | integer | | | (random() * (100)::double precision) | plain | |
  68. c63 | integer | | | (random() * (100)::double precision) | plain | |
  69. c64 | integer | | | (random() * (100)::double precision) | plain | |
  70. c65 | integer | | | (random() * (1000000)::double precision) | plain | |
  71. c66 | integer | | | (random() * (1000000)::double precision) | plain | |
  72. c67 | integer | | | (random() * (1000000)::double precision) | plain | |
  73. c68 | integer | | | (random() * (1000000)::double precision) | plain | |
  74. c69 | integer | | | (random() * (1000000)::double precision) | plain | |
  75. c70 | integer | | | (random() * (1000000)::double precision) | plain | |
  76. c71 | integer | | | (random() * (1000000)::double precision) | plain | |
  77. c72 | integer | | | (random() * (1000000)::double precision) | plain | |
  78. c73 | integer | | | (random() * (1000000)::double precision) | plain | |
  79. c74 | integer | | | (random() * (1000000)::double precision) | plain | |
  80. c75 | integer | | | (random() * (1000000)::double precision) | plain | |
  81. c76 | integer | | | (random() * (1000000)::double precision) | plain | |
  82. c77 | integer | | | (random() * (1000000)::double precision) | plain | |
  83. c78 | integer | | | (random() * (1000000)::double precision) | plain | |
  84. c79 | integer | | | (random() * (1000000)::double precision) | plain | |
  85. c80 | integer | | | (random() * (1000000)::double precision) | plain | |
  86. c81 | integer | | | (random() * (1000000)::double precision) | plain | |
  87. c82 | integer | | | (random() * (1000000)::double precision) | plain | |
  88. c83 | integer | | | (random() * (1000000)::double precision) | plain | |
  89. c84 | integer | | | (random() * (1000000)::double precision) | plain | |
  90. c85 | integer | | | (random() * (1000000)::double precision) | plain | |
  91. c86 | integer | | | (random() * (1000000)::double precision) | plain | |
  92. c87 | integer | | | (random() * (1000000)::double precision) | plain | |
  93. c88 | integer | | | (random() * (1000000)::double precision) | plain | |
  94. c89 | integer | | | (random() * (1000000)::double precision) | plain | |
  95. c90 | integer | | | (random() * (1000000)::double precision) | plain | |
  96. c91 | integer | | | (random() * (1000000)::double precision) | plain | |
  97. c92 | integer | | | (random() * (1000000)::double precision) | plain | |
  98. c93 | integer | | | (random() * (1000000)::double precision) | plain | |
  99. c94 | integer | | | (random() * (1000000)::double precision) | plain | |
  100. c95 | integer | | | (random() * (1000000)::double precision) | plain | |
  101. c96 | integer | | | (random() * (1000000)::double precision) | plain | |
  102. c97 | integer | | | (random() * (1000000)::double precision) | plain | |
  103. c98 | integer | | | (random() * (1000000)::double precision) | plain | |
  104. c99 | integer | | | (random() * (1000000)::double precision) | plain | |
  105. c100 | integer | | | (random() * (1000000)::double precision) | plain | |
  106. c101 | integer | | | (random() * (1000000)::double precision) | plain | |
  107. c102 | integer | | | (random() * (1000000)::double precision) | plain | |
  108. c103 | integer | | | (random() * (1000000)::double precision) | plain | |
  109. c104 | integer | | | (random() * (1000000)::double precision) | plain | |
  110. c105 | integer | | | (random() * (1000000)::double precision) | plain | |
  111. c106 | integer | | | (random() * (1000000)::double precision) | plain | |
  112. c107 | integer | | | (random() * (1000000)::double precision) | plain | |
  113. c108 | integer | | | (random() * (1000000)::double precision) | plain | |
  114. c109 | integer | | | (random() * (1000000)::double precision) | plain | |
  115. c110 | integer | | | (random() * (1000000)::double precision) | plain | |
  116. c111 | integer | | | (random() * (1000000)::double precision) | plain | |
  117. c112 | integer | | | (random() * (1000000)::double precision) | plain | |
  118. c113 | integer | | | (random() * (1000000)::double precision) | plain | |
  119. c114 | integer | | | (random() * (1000000)::double precision) | plain | |
  120. c115 | integer | | | (random() * (1000000)::double precision) | plain | |
  121. c116 | integer | | | (random() * (1000000)::double precision) | plain | |
  122. c117 | integer | | | (random() * (1000000)::double precision) | plain | |
  123. c118 | integer | | | (random() * (1000000)::double precision) | plain | |
  124. c119 | integer | | | (random() * (1000000)::double precision) | plain | |
  125. c120 | integer | | | (random() * (1000000)::double precision) | plain | |
  126. c121 | integer | | | (random() * (1000000)::double precision) | plain | |
  127. c122 | integer | | | (random() * (1000000)::double precision) | plain | |
  128. c123 | integer | | | (random() * (1000000)::double precision) | plain | |
  129. c124 | integer | | | (random() * (1000000)::double precision) | plain | |
  130. c125 | integer | | | (random() * (1000000)::double precision) | plain | |
  131. c126 | integer | | | (random() * (1000000)::double precision) | plain | |
  132. c127 | integer | | | (random() * (1000000)::double precision) | plain | |
  133. c128 | integer | | | (random() * (1000000)::double precision) | plain | |
  134. Indexes:
  135. "test_pkey" PRIMARY KEY, btree (id)
  136. "idx_test_1" btree (c1), tablespace "tbs_8001"
  137. "idx_test_10" btree (c10), tablespace "tbs_8001"
  138. "idx_test_100" btree (c100), tablespace "tbs_8001"
  139. "idx_test_101" btree (c101), tablespace "tbs_8001"
  140. "idx_test_102" btree (c102), tablespace "tbs_8001"
  141. "idx_test_103" btree (c103), tablespace "tbs_8001"
  142. "idx_test_104" btree (c104), tablespace "tbs_8001"
  143. "idx_test_105" btree (c105), tablespace "tbs_8001"
  144. "idx_test_106" btree (c106), tablespace "tbs_8001"
  145. "idx_test_107" btree (c107), tablespace "tbs_8001"
  146. "idx_test_108" btree (c108), tablespace "tbs_8001"
  147. "idx_test_109" btree (c109), tablespace "tbs_8001"
  148. "idx_test_11" btree (c11), tablespace "tbs_8001"
  149. "idx_test_110" btree (c110), tablespace "tbs_8001"
  150. "idx_test_111" btree (c111), tablespace "tbs_8001"
  151. "idx_test_112" btree (c112), tablespace "tbs_8001"
  152. "idx_test_113" btree (c113), tablespace "tbs_8001"
  153. "idx_test_114" btree (c114), tablespace "tbs_8001"
  154. "idx_test_115" btree (c115), tablespace "tbs_8001"
  155. "idx_test_116" btree (c116), tablespace "tbs_8001"
  156. "idx_test_117" btree (c117), tablespace "tbs_8001"
  157. "idx_test_118" btree (c118), tablespace "tbs_8001"
  158. "idx_test_119" btree (c119), tablespace "tbs_8001"
  159. "idx_test_12" btree (c12), tablespace "tbs_8001"
  160. "idx_test_120" btree (c120), tablespace "tbs_8001"
  161. "idx_test_121" btree (c121), tablespace "tbs_8001"
  162. "idx_test_122" btree (c122), tablespace "tbs_8001"
  163. "idx_test_123" btree (c123), tablespace "tbs_8001"
  164. "idx_test_124" btree (c124), tablespace "tbs_8001"
  165. "idx_test_125" btree (c125), tablespace "tbs_8001"
  166. "idx_test_126" btree (c126), tablespace "tbs_8001"
  167. "idx_test_127" btree (c127), tablespace "tbs_8001"
  168. "idx_test_128" btree (c128), tablespace "tbs_8001"
  169. "idx_test_13" btree (c13), tablespace "tbs_8001"
  170. "idx_test_14" btree (c14), tablespace "tbs_8001"
  171. "idx_test_15" btree (c15), tablespace "tbs_8001"
  172. "idx_test_16" btree (c16), tablespace "tbs_8001"
  173. "idx_test_17" btree (c17), tablespace "tbs_8001"
  174. "idx_test_18" btree (c18), tablespace "tbs_8001"
  175. "idx_test_19" btree (c19), tablespace "tbs_8001"
  176. "idx_test_2" btree (c2), tablespace "tbs_8001"
  177. "idx_test_20" btree (c20), tablespace "tbs_8001"
  178. "idx_test_21" btree (c21), tablespace "tbs_8001"
  179. "idx_test_22" btree (c22), tablespace "tbs_8001"
  180. "idx_test_23" btree (c23), tablespace "tbs_8001"
  181. "idx_test_24" btree (c24), tablespace "tbs_8001"
  182. "idx_test_25" btree (c25), tablespace "tbs_8001"
  183. "idx_test_26" btree (c26), tablespace "tbs_8001"
  184. "idx_test_27" btree (c27), tablespace "tbs_8001"
  185. "idx_test_28" btree (c28), tablespace "tbs_8001"
  186. "idx_test_29" btree (c29), tablespace "tbs_8001"
  187. "idx_test_3" btree (c3), tablespace "tbs_8001"
  188. "idx_test_30" btree (c30), tablespace "tbs_8001"
  189. "idx_test_31" btree (c31), tablespace "tbs_8001"
  190. "idx_test_32" btree (c32), tablespace "tbs_8001"
  191. "idx_test_33" btree (c33), tablespace "tbs_8001"
  192. "idx_test_34" btree (c34), tablespace "tbs_8001"
  193. "idx_test_35" btree (c35), tablespace "tbs_8001"
  194. "idx_test_36" btree (c36), tablespace "tbs_8001"
  195. "idx_test_37" btree (c37), tablespace "tbs_8001"
  196. "idx_test_38" btree (c38), tablespace "tbs_8001"
  197. "idx_test_39" btree (c39), tablespace "tbs_8001"
  198. "idx_test_4" btree (c4), tablespace "tbs_8001"
  199. "idx_test_40" btree (c40), tablespace "tbs_8001"
  200. "idx_test_41" btree (c41), tablespace "tbs_8001"
  201. "idx_test_42" btree (c42), tablespace "tbs_8001"
  202. "idx_test_43" btree (c43), tablespace "tbs_8001"
  203. "idx_test_44" btree (c44), tablespace "tbs_8001"
  204. "idx_test_45" btree (c45), tablespace "tbs_8001"
  205. "idx_test_46" btree (c46), tablespace "tbs_8001"
  206. "idx_test_47" btree (c47), tablespace "tbs_8001"
  207. "idx_test_48" btree (c48), tablespace "tbs_8001"
  208. "idx_test_49" btree (c49), tablespace "tbs_8001"
  209. "idx_test_5" btree (c5), tablespace "tbs_8001"
  210. "idx_test_50" btree (c50), tablespace "tbs_8001"
  211. "idx_test_51" btree (c51), tablespace "tbs_8001"
  212. "idx_test_52" btree (c52), tablespace "tbs_8001"
  213. "idx_test_53" btree (c53), tablespace "tbs_8001"
  214. "idx_test_54" btree (c54), tablespace "tbs_8001"
  215. "idx_test_55" btree (c55), tablespace "tbs_8001"
  216. "idx_test_56" btree (c56), tablespace "tbs_8001"
  217. "idx_test_57" btree (c57), tablespace "tbs_8001"
  218. "idx_test_58" btree (c58), tablespace "tbs_8001"
  219. "idx_test_59" btree (c59), tablespace "tbs_8001"
  220. "idx_test_6" btree (c6), tablespace "tbs_8001"
  221. "idx_test_60" btree (c60), tablespace "tbs_8001"
  222. "idx_test_61" btree (c61), tablespace "tbs_8001"
  223. "idx_test_62" btree (c62), tablespace "tbs_8001"
  224. "idx_test_63" btree (c63), tablespace "tbs_8001"
  225. "idx_test_64" btree (c64), tablespace "tbs_8001"
  226. "idx_test_65" btree (c65), tablespace "tbs_8001"
  227. "idx_test_66" btree (c66), tablespace "tbs_8001"
  228. "idx_test_67" btree (c67), tablespace "tbs_8001"
  229. "idx_test_68" btree (c68), tablespace "tbs_8001"
  230. "idx_test_69" btree (c69), tablespace "tbs_8001"
  231. "idx_test_7" btree (c7), tablespace "tbs_8001"
  232. "idx_test_70" btree (c70), tablespace "tbs_8001"
  233. "idx_test_71" btree (c71), tablespace "tbs_8001"
  234. "idx_test_72" btree (c72), tablespace "tbs_8001"
  235. "idx_test_73" btree (c73), tablespace "tbs_8001"
  236. "idx_test_74" btree (c74), tablespace "tbs_8001"
  237. "idx_test_75" btree (c75), tablespace "tbs_8001"
  238. "idx_test_76" btree (c76), tablespace "tbs_8001"
  239. "idx_test_77" btree (c77), tablespace "tbs_8001"
  240. "idx_test_78" btree (c78), tablespace "tbs_8001"
  241. "idx_test_79" btree (c79), tablespace "tbs_8001"
  242. "idx_test_8" btree (c8), tablespace "tbs_8001"
  243. "idx_test_80" btree (c80), tablespace "tbs_8001"
  244. "idx_test_81" btree (c81), tablespace "tbs_8001"
  245. "idx_test_82" btree (c82), tablespace "tbs_8001"
  246. "idx_test_83" btree (c83), tablespace "tbs_8001"
  247. "idx_test_84" btree (c84), tablespace "tbs_8001"
  248. "idx_test_85" btree (c85), tablespace "tbs_8001"
  249. "idx_test_86" btree (c86), tablespace "tbs_8001"
  250. "idx_test_87" btree (c87), tablespace "tbs_8001"
  251. "idx_test_88" btree (c88), tablespace "tbs_8001"
  252. "idx_test_89" btree (c89), tablespace "tbs_8001"
  253. "idx_test_9" btree (c9), tablespace "tbs_8001"
  254. "idx_test_90" btree (c90), tablespace "tbs_8001"
  255. "idx_test_91" btree (c91), tablespace "tbs_8001"
  256. "idx_test_92" btree (c92), tablespace "tbs_8001"
  257. "idx_test_93" btree (c93), tablespace "tbs_8001"
  258. "idx_test_94" btree (c94), tablespace "tbs_8001"
  259. "idx_test_95" btree (c95), tablespace "tbs_8001"
  260. "idx_test_96" btree (c96), tablespace "tbs_8001"
  261. "idx_test_97" btree (c97), tablespace "tbs_8001"
  262. "idx_test_98" btree (c98), tablespace "tbs_8001"
  263. "idx_test_99" btree (c99), tablespace "tbs_8001"
  264. Options: parallel_workers=64

写入性能如何

当前有129个索引,写入性能如何?

9505行/s。

  1. transaction type: ./test.sql
  2. scaling factor: 1
  3. query mode: prepared
  4. number of clients: 24
  5. number of threads: 24
  6. duration: 120 s
  7. number of transactions actually processed: 11433
  8. latency average = 252.195 ms
  9. latency stddev = 70.089 ms
  10. tps = 95.054689 (including connections establishing)
  11. tps = 95.058210 (excluding connections establishing)
  12. statement latencies in milliseconds:
  13. 252.179 insert into test (c1) select random()*100 from generate_series(1,100);

瓶颈,磁盘读写5.5GB/s。

  1. Total DISK READ : 207.91 K/s | Total DISK WRITE : 3.54 G/s
  2. Actual DISK READ: 207.91 K/s | Actual DISK WRITE: 2015.64 M/s
  3. TID PRIO USER DISK READ DISK WRITE SWAPIN IO> COMMAND
  4. 55887 be/4 digoal 15.40 K/s 158.54 M/s 0.00 % 1.05 % postgres: postgres postgres [local] INSERT
  5. 55872 be/4 digoal 7.70 K/s 157.62 M/s 0.00 % 0.84 % postgres: postgres postgres [local] INSERT
  6. 55886 be/4 digoal 23.10 K/s 158.78 M/s 0.00 % 0.78 % postgres: postgres postgres [local] INSERT
  7. 55897 be/4 digoal 7.70 K/s 158.79 M/s 0.00 % 0.75 % postgres: postgres postgres [local] INSERT
  8. 55889 be/4 digoal 0.00 B/s 158.72 M/s 0.00 % 0.69 % postgres: postgres postgres [local] INSERT
  9. 55894 be/4 digoal 0.00 B/s 157.25 M/s 0.00 % 0.69 % postgres: postgres postgres [local] INSERT
  10. 55888 be/4 digoal 7.70 K/s 136.26 M/s 0.00 % 0.68 % postgres: postgres postgres [local] INSERT
  11. 55885 be/4 digoal 7.70 K/s 143.24 M/s 0.00 % 0.67 % postgres: postgres postgres [local] INSERT
  12. 55890 be/4 digoal 0.00 B/s 159.07 M/s 0.00 % 0.67 % postgres: postgres postgres [local] INSERT
  13. 55865 be/4 digoal 15.40 K/s 158.27 M/s 0.00 % 0.65 % postgres: postgres postgres [local] INSERT
  14. 55900 be/4 digoal 7.70 K/s 151.00 M/s 0.00 % 0.64 % postgres: postgres postgres [local] INSERT
  15. 55891 be/4 digoal 0.00 B/s 160.40 M/s 0.00 % 0.63 % postgres: postgres postgres [local] INSERT
  16. 55896 be/4 digoal 0.00 B/s 158.79 M/s 0.00 % 0.62 % postgres: postgres postgres [local] INSERT
  17. 55902 be/4 digoal 15.40 K/s 157.65 M/s 0.00 % 0.62 % postgres: postgres postgres [local] INSERT
  18. 55875 be/4 digoal 0.00 B/s 158.52 M/s 0.00 % 0.58 % postgres: postgres postgres [local] INSERT
  19. 55892 be/4 digoal 7.70 K/s 136.20 M/s 0.00 % 0.58 % postgres: postgres postgres [local] INSERT
  20. 55868 be/4 digoal 0.00 B/s 139.10 M/s 0.00 % 0.58 % postgres: postgres postgres [local] INSERT
  21. 55895 be/4 digoal 0.00 B/s 159.75 M/s 0.00 % 0.57 % postgres: postgres postgres [local] INSERT
  22. 55898 be/4 digoal 0.00 B/s 113.43 M/s 0.00 % 0.55 % postgres: postgres postgres [local] INSERT
  23. 55880 be/4 digoal 46.20 K/s 121.68 M/s 0.00 % 0.50 % postgres: postgres postgres [local] INSERT
  24. 55884 be/4 digoal 23.10 K/s 126.35 M/s 0.00 % 0.47 % postgres: postgres postgres [local] INSERT
  25. 55901 be/4 digoal 15.40 K/s 117.46 M/s 0.00 % 0.46 % postgres: postgres postgres [local] INSERT
  26. 55899 be/4 digoal 7.70 K/s 115.13 M/s 0.00 % 0.46 % postgres: postgres postgres [local] INSERT

瓶颈在读写数据文件

  1. postgres=# select wait_event_type,wait_event,count(*) from pg_stat_activity where wait_event is not null group by 1,2 order by 3 desc;
  2. wait_event_type | wait_event | count
  3. -----------------+---------------------+-------
  4. IO | DataFileWrite | 15
  5. IO | DataFileRead | 5
  6. Activity | WalWriterMain | 1
  7. Activity | LogicalLauncherMain | 1
  8. Activity | CheckpointerMain | 1
  9. Activity | AutoVacuumMain | 1
  10. (6 rows)

任意字段组合查询性能如何

1、

  1. postgres=# explain select count(*) from test where c1=2 and c99 between 100 and 1000 and c98 between 100 and 200 and c1=1;
  2. QUERY PLAN
  3. ---------------------------------------------------------------------------------------------------
  4. Aggregate (cost=1201.23..1201.24 rows=1 width=8)
  5. -> Result (cost=1192.25..1201.22 rows=1 width=0)
  6. One-Time Filter: false
  7. -> Bitmap Heap Scan on test (cost=1192.25..1201.22 rows=1 width=0)
  8. Recheck Cond: ((c98 >= 100) AND (c98 <= 200) AND (c99 >= 100) AND (c99 <= 1000))
  9. Filter: (c1 = 2)
  10. -> BitmapAnd (cost=1192.25..1192.25 rows=8 width=0)
  11. -> Bitmap Index Scan on idx_test_98 (cost=0.00..125.98 rows=9571 width=0)
  12. Index Cond: ((c98 >= 100) AND (c98 <= 200))
  13. -> Bitmap Index Scan on idx_test_99 (cost=0.00..1066.02 rows=81795 width=0)
  14. Index Cond: ((c99 >= 100) AND (c99 <= 1000))
  15. (11 rows)
  16. postgres=# select count(*) from test where c1=2 and c99 between 100 and 1000 and c98 between 100 and 200 and c2=1;
  17. count
  18. -------
  19. 0
  20. (1 row)
  21. Time: 1.087 ms

2、

  1. set min_parallel_index_scan_size=0;
  2. set min_parallel_table_scan_size=0;
  3. set parallel_setup_cost=0;
  4. set parallel_tuple_cost=0;
  5. set work_mem='1GB';
  6. set max_parallel_workers=128;
  7. set max_parallel_workers_per_gather=24;
  8. set random_page_cost =1.1;
  9. set effective_cache_size ='400GB';
  10. alter table test set (parallel_workers=64);
  11. set enable_bitmapscan=off;
  1. postgres=# select count(*) from test where c1=2 and c99 between 100 and 10000;
  2. count
  3. -------
  4. 9764
  5. (1 row)
  6. Time: 50.160 ms
  7. postgres=# select count(*) from test where c1=2 and c99 between 100 and 1000 and c98 between 100 and 200 and c2=1;
  8. count
  9. -------
  10. 0
  11. (1 row)
  12. Time: 20.969 ms
  13. postgres=# select count(*) from test where c1=2 and c99 between 100 and 10000 and c108 between 100 and 10000;
  14. count
  15. -------
  16. 102
  17. (1 row)
  18. Time: 72.359 ms
  19. postgres=# select count(*) from test where c1=2 and c99=1;
  20. count
  21. -------
  22. 2
  23. (1 row)
  24. Time: 1.118 ms

3、OR

  1. set enable_bitmapscan=on;
  2. postgres=# explain select count(*) from test where c1=2 and c99=1 or c100 between 10 and 100;
  3. QUERY PLAN
  4. --------------------------------------------------------------------------------------------
  5. Aggregate (cost=10000010781.91..10000010781.92 rows=1 width=8)
  6. -> Bitmap Heap Scan on test (cost=10000000130.57..10000010758.33 rows=9430 width=0)
  7. Recheck Cond: ((c99 = 1) OR ((c100 >= 10) AND (c100 <= 100)))
  8. Filter: (((c1 = 2) AND (c99 = 1)) OR ((c100 >= 10) AND (c100 <= 100)))
  9. -> BitmapOr (cost=130.57..130.57 rows=9526 width=0)
  10. -> Bitmap Index Scan on idx_test_99 (cost=0.00..2.39 rows=96 width=0)
  11. Index Cond: (c99 = 1)
  12. -> Bitmap Index Scan on idx_test_100 (cost=0.00..123.47 rows=9430 width=0)
  13. Index Cond: ((c100 >= 10) AND (c100 <= 100))
  14. (9 rows)
  15. Time: 1.281 ms
  16. postgres=# select count(*) from test where c1=2 and c99=1 or c100 between 10 and 100;
  17. count
  18. -------
  19. 9174
  20. (1 row)
  21. Time: 18.785 ms

小结

任意维度查询case耗时
c1=2 and c99 between 100 and 10000;50 毫秒
c1=2 and c99 between 100 and 1000 and c98 between 100 and 200 and c2=1;21 毫秒
c1=2 and c99 between 100 and 10000 and c108 between 100 and 10000;72 毫秒
c1=2 and c99=1;1 毫秒
c1=2 and c99=1 or c100 between 10 and 100;19 毫秒

性能差异:

1、执行计划

2、扫描量

3、运算量(与结果集大小无直接关系,关键看扫描方法和中间计算量)。

写入能力:129个索引,写入9505行/s。瓶颈在IO侧,通过提升IO能力,加分区可以提高。

参考

《PostgreSQL 设计优化case - 大宽表任意字段组合查询索引如何选择(btree, gin, rum) - (含单个索引列数超过32列的方法)》

《PostgreSQL 任意字段数组合 AND\OR 条件,指定返回结果条数,构造测试数据算法举例》

《PostgreSQL ADHoc(任意字段组合)查询(rums索引加速) - 非字典化,普通、数组等组合字段生成新数组》

《PostgreSQL 实践 - 实时广告位推荐 2 (任意字段组合、任意维度组合搜索、输出TOP-K)》

《PostgreSQL 实践 - 实时广告位推荐 1 (任意字段组合、任意维度组合搜索、输出TOP-K)》

《PostgreSQL ADHoc(任意字段组合)查询 与 字典化 (rum索引加速) - 实践与方案1》

《PostgreSQL 如何高效解决 按任意字段分词检索的问题 - case 1》

《HTAP数据库 PostgreSQL 场景与性能测试之 20 - (OLAP) 用户画像圈人场景 - 多个字段任意组合条件筛选与透视》

《PostgreSQL 多字段任意组合搜索的性能》

《PostgreSQL 并行计算解说 汇总》

《PostgreSQL 9种索引的原理和应用场景》

原文:http://mysql.taobao.org/monthly/2019/04/09/