Postgres select 大型（1500 万行）table 极慢，即使有索引

Question

我正在尝试运行 EXPLAIN ANALYZE 但它根本无法完成，因为它太慢了。如果是这样，我会 post 结果，但现在，这里是 EXPLAIN.

查询：

EXPLAIN SELECT
    *
FROM
    "Posts" AS "Post"
WHERE
    (
        "Post"."featurePostOnDate" > '2020-06-25 19:28:07.816 +00:00'
        OR (
            "Post"."featurePostOnDate" IS NULL
            AND "Post"."userId" IN (6863684)
        )
    )
AND "Post"."private" IS NULL
ORDER BY
    "Post"."featurePostOnDate" DESC NULLS LAST,
    "Post"."createdAt" DESC NULLS LAST
LIMIT 10;

结果：

Limit  (cost=0.56..110.92 rows=10 width=1136)
  ->  Index Scan using posts_updated_following_feed_idx on "Posts" "Post"  (cost=0.56..284949.60 rows=25819 width=1136)
        Filter: (("featurePostOnDate" > '2020-06-25 19:28:07.816+00'::timestamp with time zone) OR (("featurePostOnDate" IS NULL) AND ("userId" = 6863684)))

索引：

CREATE INDEX  "posts_updated_following_feed_idx" ON "public"."Posts" USING btree (
    "featurePostOnDate" DESC NULLS LAST,
    "createdAt" DESC NULLS LAST
)
WHERE
    private IS NULL;

Answer 1

因此，由于您有 1500 万行，并且您使用了 ANALYZE。使用 ANALYZE 实际运行查询，您可以从这里引用它 https://www.postgresql.org/docs/9.1/sql-explain.html.

并且在 WHERE 子句中您使用了未编入索引的字段

WHERE
    (
        "Post"."featurePostOnDate" > '2020-06-25 19:28:07.816 +00:00'
        OR (
            "Post"."featurePostOnDate" IS NULL
            AND "Post"."userId" IN (6863684)
        )
    )
AND "Post"."private" IS NULL

所以它实际上是在进行顺序扫描以过滤掉行

Filter: (("featurePostOnDate" > '2020-06-25 19:28:07.816+00'::timestamp with time zone) OR (("featurePostOnDate" IS NULL) AND ("userId" = 6863684)))

这可能是您的查询速度慢的原因。

您可能需要 (featurePostOnDate, userId, private) 和 (featurePostOnDate, private).

上的复合索引

希望对您有所帮助。

Answer 2

您需要将其编写为两个单独的查询，一个用于 OR 的每个分支。将限制应用于每个查询，然后组合它们并再次联合应用限制。但是，如果第一个分支找到十行，则第二个分支根本不需要运行，因为所有非 NULL 日期都已经排在第一位。

Postgres select 大型（1500 万行）table 极慢，即使有索引

Postgres select on large (15m rows) table extremely slow, even with index

postgresql

indexing

database-performance