Question

我有以下postgres查询，该查询可为我提供所需的结果，但由于运行两次相同的内部查询，我觉得它没有经过优化。我尝试了一个分组依据，但一直说我必须将xdata.column_name添加到分组依据。我必须获取total_count，因为我正在使用这些结果创建分页。任何帮助将不胜感激。

SELECT total.total_count, xdata.* FROM (SELECT * FROM (SELECT *, CAST(celery_taskresult.meta -> 'args' ->> 0 AS INT) AS user_id, celery_taskresult.meta ->> 'name' AS task_name FROM celery_taskresult) AS x 
    WHERE x.user_id IS NOT NULL) as xdata
LEFT JOIN (SELECT COUNT(z.id) AS total_count, 0 AS b FROM (SELECT *, CAST(celery_taskresult.meta -> 'args' ->> 0 AS INT) AS user_id, celery_taskresult.meta ->> 'name' AS task_name FROM celery_taskresult) AS z 
    WHERE z.user_id IS NOT NULL) as total ON 0 = total.b
ORDER BY xdata.id ASC
LIMIT 25 OFFSET 0

Answer 1

我已经有一段时间没有使用postgresql了，但是选择COUNT(*) OVER()应该可以

  SELECT COUNT(*) OVER() AS total_count, 
         xdata.* 
    FROM (SELECT * 
            FROM (SELECT *, 
                         CAST(celery_taskresult.meta -> 'args' ->> 0 AS INT) AS user_id, 
                         celery_taskresult.meta ->> 'name' AS task_name 
                    FROM celery_taskresult) AS X 
           WHERE x.user_id IS NOT NULL) AS xdata
ORDER BY xdata.id ASC
LIMIT 25 OFFSET 0

编辑：PostgreSQL窗口上的文档窗口功能-Window Functions

Answer 2

您可以使用cte。 https://www.tutorialspoint.com/postgresql/postgresql_with_clause.htm。例如
使用cte作为（从a中选择*），从cte中选择*；
将内部的select *替换为select *，并使用它。

通过减少内部查询来优化此SQL查询？

2 个答案: