首页 > 解决方案 > POSTGRES - 子查询返回结果非常慢

问题描述

我有一个相当简单的查询,但它包含一个子查询。我想按顺序获取resource_id通过过滤器的列表。ORDER BY DESC;

更多信息 基本上我们需要找到resource_id's通过下面查询中提到的过滤器的列表,

   SELECT rs.resource_id
FROM resource rs
WHERE (
    SELECT rc.resource_id
    FROM risk_child rc
    WHERE rc.resource_id = rs.resource_id 
    AND rc.cloudaccount_id = rs.cloud_account_id
    AND rs.reg_id= any(array[236]) 
    AND rc.risk_level= any(array['high','low'])
    AND rc.status = any(array['fail'])
    AND rc.cloudaccount_id= any (array['4ZiCmwslbjhmRtHAOjLG'])
    ORDER BY rc.id DESC
    LIMIT 1
) = rs.resource_id

然后这些资源将被传递到另一个查询中,如此处所述:

SELECT
  DISTINCT ON (rc.resource_id, rc.rule_id, s.id) MAX(rc.creationtime) as creationtime,
  rc.resource_id,
  rl.rule_tag,
  s.service,
  r.region,
  rc.status,
  rs.vpc_id,
  rc.cloudaccount_id,
  rc.organization_id,
  rs.owner_id,
  rc.description,
  f.function_name,
  g.group_name,
  rc.risk_level,
  rc.id,
  rc.user_id,
  rc.pro_id,
  c.category_name,
  rc.raw as rawResponse,
  rs.res_ca_id,
  rs.resource_name
FROM
  risk_child rc,
  resource rs,
  rule rl,
  service s,
  region r,
  function f,
  g_by g,
  category c
WHERE
  rc.resource_id = rs.resource_id
  AND rl.id = rc.rule_id
  AND s.id = rs.ser_id
  AND rs.reg_id = r.id
  AND f.id = rc.function_id
  AND c.id = rc.category_id
  AND g.id = rc.group_id
  AND rc.cloudaccount_id like any (array $ { modifiedCloudAccounts })
  AND rc.organization_id = $ { orgId }
  AND rc.rule_id > 0
  AND rc.cloudaccount_id = rs.cloud_account_id
  AND rs.resource_id like any (array $ { getResources }) $ { risk }
GROUP BY
  rc.rule_id,
  rc.creationtime,
  rc.creationtime,
  rc.resource_id,
  rl.rule_tag,
  rl.id,
  s.service,
  r.region,
  rc.status,
  rs.vpc_id,
  rc.cloudaccount_id,
  rc.organization_id,
  rs.owner_id,
  rc.description,
  f.function_name,
  g.group_name,
  rc.risk_level,
  rc.id,
  rc.user_id,
  rc.pro_id,
  c.category_name,
  rc.raw,
  s.id,
  rs.res_ca_id
ORDER BY
  rc.resource_id,
  rc.rule_id ASC;

问题 现在第一个查询返回结果非常慢,即使在索引结果后 5 - 6 秒。所以请记住第一个查询需要运行两次

  1. 一获取总行数(用于分页)
  2. 第二次获取resource_ids

我主要在我的应用程序中使用 NO - SQL,所以我对 SQL 查询相当陌生。任何帮助,将不胜感激。谢谢

标签: sqlpostgresqlsubquery

解决方案


所以最后这件事适用于我的情况:

    SELECT rs.resource_id
FROM resource rs
WHERE EXISTS (SELECT *
              FROM risk_child rc
              WHERE rc.resource_id = rs.resource_id 
                AND rc.cloudaccount_id = rs.cloud_account_id
                AND rs.reg_id= any(array[236]) 
                AND rc.risk_level= any(array['high','low'])
                AND rc.status = any(array['fail'])
                AND rc.cloudaccount_id= any (array['4ZiCmwslbjhmRtHAOjLG'])
             )

基本上,正如我在帖子中解释的那样,我对 SQL 相当陌生,我没有在我的表中放置正确的索引,所以我必须添加以下索引以使查询更快

resource(resource_id, cloud_account_id)risk_child(resource_id, cloudaccount_id)这帮助我进一步提高了我的表现。


推荐阅读