Question

假设我有以下（2）表：

[人员表]

ID 名称
约翰一 2 Suzy
3戴夫

[请求表]

ID 1stChoicePersonRequested 2ndChoicePersonRequested 3rdChoicePersonRequested
1 1 2 3
2 2 NULL NULL
3 1 2 NULL
4 3 1 1

我在查询中寻找的结果如下：

NameID 名称 NumTimeRequested
约翰一书4 2 Suzy 3
3戴夫2

我需要结果来计算每个人被请求的总次数，而不管它们包含哪些（3）列。在SQL 2008中执行此操作的最佳方法是什么？谢谢！

Answer 1

SELECT P.ID NameID, Name, C NumTimeRequested
FROM
    (
        SELECT ID, COUNT(*) C
        FROM (
            SELECT [1stChoicePersonRequested] ID FROM [Request Table]
            UNION ALL
            SELECT [2ndChoicePersonRequested] ID FROM [Request Table]
            UNION ALL
            SELECT [3rdChoicePersonRequested] ID FROM [Request Table]
        ) R1
        GROUP BY ID
    ) R2
    JOIN [People Table] P ON R2.ID = P.ID

或者，你可以这样做：

SELECT
    ID NameID,
    Name,
    (SELECT COUNT(*) FROM [Request Table] WHERE [1stChoicePersonRequested] = P.ID)
    + (SELECT COUNT(*) FROM [Request Table] WHERE [2ndChoicePersonRequested] = P.ID)
    + (SELECT COUNT(*) FROM [Request Table] WHERE [3rdChoicePersonRequested] = P.ID)    
    NumTimeRequested
FROM
    [People Table] P

Answer 2

;with R as
(
  select PersonID
  from [Request Table]
    cross apply (values ([1stChoicePersonRequested]),
                        ([2ndChoicePersonRequested]),
                        ([3rdChoicePersonRequested])) as T(PersonID)
)
select R.PersonID as NameID, 
       P.Name, 
       count(*) as NumTimeRequested 
from R
  inner join [People Table] as P
    on R.PersonID = P.ID
group by R.PersonID, P.Name
order by R.PersonID

http://data.stackexchange.com/stackoverflow/q/122030/

如何统计可能跨越SQL中多列的ID？

2 个答案: