SQL加入查詢如何到

Question

CREATE TABLE interview (uniqueID int identity(1,1), 
         date datetime, 
         recordtype int, 
         amount numeric(18, 4)) 

INSERT INTO interview values('6/30/13', 1, 27.95) 
INSERT INTO interview values('5/20/13', 1, 21.85) 
INSERT INTO interview values('5/22/13', 2, 27.90) 
INSERT INTO interview values('12/11/12', 2, 23.95) 
INSERT INTO interview values('6/13/13', 3, 24.90) 
INSERT INTO interview values('6/30/13', 2, 27.95) 
INSERT INTO interview values('5/20/13', 2, 21.85) 
INSERT INTO interview values('5/22/13', 1, 27.90) 
INSERT INTO interview values('12/11/12',1, 23.95) 
INSERT INTO interview values('6/13/13', 3, 24.90) 
INSERT INTO interview values('6/30/13', 3, 27.95) 
INSERT INTO interview values('5/20/13', 3, 21.85) 
INSERT INTO interview values('5/22/13', 2, 27.90) 
INSERT INTO interview values('12/11/12', 1, 23.95) 
INSERT INTO interview values('6/13/13', 1, 24.90) 

如何獲得以下結果？查詢會是什麼樣子？

我只能得到部分的工作，但我的答案是不正確的。我需要 以某種方式加入查詢。

select distinct date, count(RecordType)as Count_unique1 
from interview 
where RecordType = '1' 
group by date 

select distinct date, count(RecordType)as Count_unique2 
from interview 
where RecordType = '2' 
group by date 

select distinct date, count(RecordType)as Count_unique3 
from interview 
where RecordType = '3' 
group by date 

Answer 1

select 
date, 
sum(case when RecordType = '1' then 1 else 0 end) as Count_unique1, 
sum(case when RecordType = '2' then 1 else 0 end) as Count_unique2, 
sum(case when RecordType = '3' then 1 else 0 end) as Count_unique3 
from interview 
group by date 

Answer 2

如果記錄類型是在所有的情況下，這將是足夠的1,2和3。

select date, 
sum(RecordType = '1') as Count_unique1, 
sum(RecordType = '2') as Count_unique2, 
sum(RecordType = '3') as Count_unique3 
from interview 
group by date 

Answer 3

另外在MSSQL你可以使用PIVOT

SELECT date, [1] AS Count_unique1, 
      [2] AS Count_unique2, 
      [3] AS Count_unique3 
FROM (SELECT date,recordtype,amount FROM interview) p 
PIVOT 
(
    COUNT (amount) 
    FOR recordtype IN ([1], [2], [3]) 
) AS pvt 
ORDER BY pvt.date; 

SQLFiddle demo