Question

我正在编写一个SQL查询来计算该特定日期的用户总数

查询如下： -

select plan_date, type_name, account_type_id, new_type
FROM (
    select
        final_reporting_db.plan_change_facts.date as plan_date,
        final_reporting_db.plan_dim.account_type as type_name,
        final_reporting_db.plan_dim.id as account_type_id,
        (
            select count(*) as new_type from final_reporting_db.plan_change_facts
            where final_reporting_db.plan_change_facts.plan_status_dim_id = account_type_id
            and final_reporting_db.plan_change_facts.date = plan_date
        ) as new_type
    FROM 
        final_reporting_db.plan_change_facts
    INNER JOIN final_reporting_db.plan_dim where final_reporting_db.plan_dim.id in(1,2,3,4,5,6,8,9,13)
) main_table where plan_date between '2016-04-24' and '2016-04-28' group by plan_date, type_name;

在此查询子查询中花费了太多时间来执行，因为它正在计算该日期的计数。

我对如何提高此查询性能感到困惑。因为执行需要36秒。有没有办法使用group by子句而不是使用嵌套查询。帮助将不胜感激

Answer 1

评论太长了。

首先，您的查询几乎无法阅读。我提出的最好的是：

select cf.date as plan_date,
       d.account_type as type_name, d.id as account_type_id,
       (select count(*) as new_type
        from final_reporting_db.plan_change_facts cf2
        where cf2.plan_status_dim_id = ??.account_type_id and 
              cf2.date = ??.plan_date
-------------------------^ I assume this should be cf
       ) as new_type
from final_reporting_db.plan_change_facts cf join
     final_reporting_db.plan_dim d
-----^ where is the `on` clause ???
where d.id in(1,2,3,4,5,6,8,9,13) and
      cf.date between '2016-04-24' and '2016-04-28'
group by plan_date, type_name;

我观察到以下内容：

子查询中的关联条件无效，因为它们只比较子查询中的列。
您缺少最终加入的on子句，因此这是笛卡尔积。
我怀疑你使用的是MySQL，它实现了子查询。因此，您不希望在group by。

Nestesed SQL查询需要花费很多时间来执行

1 个答案: