Question

似乎我创建的查询有点乱。还有其他方法可以让它更具可读性吗？

这是我创建的查询。

Select nullif((select count(*) from [casino_game].[dbo].[group_user] gu
        where gu.group_id = (select pt.group_id from [casino_game].[dbo].[promo_trigger] pt
        join [casino_game].[dbo].[promo_offer] po on po.trigger_id = pt.trigger_id
        where pt.name = f.promo_name
        group by pt.group_id)),0) as target_group,                           -- Number of customer uploaded to a free game program.

        (cast(count(*) as decimal(11,2)) / nullif((select count(*) from [casino_game].[dbo].[group_user] gu
        where gu.group_id = (select pt.group_id from [casino_game].[dbo].[promo_trigger] pt
        join [casino_game].[dbo].[promo_offer] po on po.trigger_id = pt.trigger_id
        where pt.name = f.promo_name
        group by pt.group_id)), 0)) * 100 as claim_rate from [data].[dbo].[testgame]

这个查询正在运行，但是我需要一些建议，如果有另一种方法可以使这个工作更具可读性。

Answer 1

您可以使用In [628]: %timeit df.loc[np.in1d(df['id'], ['0','1']),'id'].map({'0': False, '1': True}) 100 loops, best of 3: 2.19 ms per loop来避免使用相同的OUTER APPLY两次

correlated sub-query

Answer 2

我认为目标表是group_user而不是testgame，它必须是一行，不需要连接到testgame的多行

select 
    COUNT(*) as target_group,  -- Number of customer uploaded to a free game program.
    -- It may cause the Divide by zero error if count(*) = 0
    CASE WHEN COUNT(*) > 0 THEN (cast((SELECT COUNT(*) as value FROM [data].[dbo].[testgame]) as decimal(11,2)) / COUNT(*)) * 100 END as claim_rate 
from 
    [casino_game].[dbo].[group_user] gu 
    inner join [casino_game].[dbo].[promo_trigger] pt on gu.group_id = pt.group_id
    inner join [casino_game].[dbo].[promo_offer] po on po.trigger_id = pt.trigger_id

从Select Query中的另一个表中选择Columns

2 个答案: