有时Postgres中的查询计数缓慢

Question

通常在一段时间内第一次会花费很多时间。在那之后，将花费相似的时间。

表大小：6.4M

explain analyze
SELECT count(*) as "count", "bookId" 
FROM "libraries"
WHERE "bookId" IN ('1', '2', '3', '4', '5')
GROUP BY "bookId";

执行时间

• 第一轮：4500ms
• 第二轮：105ms
• 第三轮：92ms
• 第四轮：93ms

解释分析

GroupAggregate  (cost=0.56..17520.53 rows=4308 width=33) (actual time=438.948..2258.747 rows=9 loops=1)
  Group Key: "bookId"
  ->  Index Only Scan using "libraries_bookId" on libraries  (cost=0.56..17445.19 rows=6452 width=25) (actual time=0.798..2249.024 rows=32306 loops=1)
        Index Cond: ("bookId" = ANY ('{1, 2, 3, 4, 5}'::text[]))
        Heap Fetches: 32306
Planning time: 0.254 ms
Execution time: 2258.813 ms