Question

我有一个名为Campaign的表。看起来像这样：

Bj::value

我想通过事件类型列中的不同值对其进行分组，然后计算事件类型中的不同值。我的桌子应该看起来像这样：

j > i

我该怎么做？

编辑14/03/19：

到目前为止我要去的地方：

website event type
amazon  imp
amazon  imp
amazon  imp
amazon  click
apple   click
apple   imp
adidas  click
adidas  imp
adidas  click
adidas  imp

Answer 1

您可以将PIVOT与COUNT一起使用

DECLARE @SampleTable TABLE (website VARCHAR(10), [event type]  VARCHAR(10))
INSERT INTO @SampleTable VALUES
('amazon' ,'imp'),
('amazon' ,'imp'),
('amazon' ,'imp'),
('amazon' ,'click'),
('apple' ,'click'),
('apple' ,'imp'),
('adidas' ,'click'),
('adidas' ,'imp'),
('adidas' ,'click'),
('adidas' ,'imp')


SELECT * FROM @SampleTable
PIVOT (COUNT([event type]) FOR [event type] IN ([imp], [click])) PVT

结果：

website    imp         click
---------- ----------- -----------
adidas     2           2
amazon     3           1
apple      1           1

Answer 2

您需要条件聚合：

select website,
       sum(case when event = 'imp' then 1 else 0 end) as imp,
       sum(case when event = 'click' then 1 else 0 end) as click
from campaign c
group by website;

Answer 3

在其他用户回答的可用其他选项中，您也可以尝试以下查询。

select a.website, imp, click from(
     select website, count(eventtype) as imp
     from test a where a.eventtype = 'imp' group by website
)a 
inner join (
     select website, count(eventtype) as click
     from test b where b.eventtype = 'click' group by website  
)b on a.website = b.website

Demo

如何同时对具有2个值的列进行分组和计数？

3 个答案: