使用 MS SQL Server 中的 Group by 字段将垂直 sql 结果转换为水平输出

Question

我有一个包含垂直销售交易数据的表（在 SQL Server 2008 中），我想将垂直输出转换为水平输出，我尝试使用 PIVOT，但有些人不知道如何在 PIVOT 上应用 group by ，因为我想要基于 AccountHeadID、TransType 和 IsPast Column 的 Sum。

示例表

CREATE TABLE Trans
    ([TransID] int,
     [CustID] int,
     [AccountHeadID] int,
     [TransType] varchar(100),
     [IsPast] bit,
     [Amount] int)
;

INSERT INTO Trans
    ([TransID],CustID, [AccountHeadID], [TransType], [IsPast],[Amount])
VALUES
    (1, 1, 1, 'Sales',1,1000),
    (2, 1, 1, 'Sales',0,500),
    (3, 1, 1, 'Sales',0,400),
    (4, 1, 1, 'Return',0,300),
    (5, 1, 1, 'Return',0,200),
    (6, 1, 1, 'Return',0,100),
    (7, 1, 1, 'Return',1,150),
    (8, 1, 2, 'Sales',1,2000)   

当前查询

DECLARE @cols AS NVARCHAR(MAX),
    @query  AS NVARCHAR(MAX)

select @cols = STUFF((SELECT ',' + QUOTENAME(AccountHeadID)  
                    from Trans
                    group by AccountHeadID
                    order by AccountHeadID
            FOR XML PATH(''), TYPE
            ).value('.', 'NVARCHAR(MAX)') 
        ,1,1,'')


set @query = 'SELECT CustID,' + @cols + ' 
            from 
            (
              SELECT 
                a.CustID,
                a.AccountHeadID,
                 a.TransType,
                 a.Amount,
                 a.isPast
              FROM Trans a
            ) x
            pivot 
            (
                sum(Amount)
                for AccountHeadID in (' + @cols + ')
            ) p '

execute sp_executesql @query;

预期输出

CustID | Account-HeadID-TransType-IsPast[1-Sales-Past] |  Account-HeadID-TransType-IsCurrent[1-Sales-Current] | Account-HeadID-TransType-IsCurrent[1-Return-Past] |  Account-HeadID-TransType-IsCurrent[1-Return-Current] | Account-HeadID-TransType-IsCurrent[2-Sales-Past]| ...
1      |1000                                           | 900 (500 + 400)                                      |150                                                | 600[300+200+100]                                       |2000   

见SQL Fiddle with Demo

欢迎提出任何建议或意见！

谢谢 苏雷什

Answer 1

试试这个：

DECLARE @cols AS NVARCHAR(MAX),
    @query  AS NVARCHAR(MAX)

select @cols = STUFF((SELECT ',' + QUOTENAME(CONVERT(NVARCHAR(10),AccountHeadID) + N'-' + TransType + N'-' + CONVERT(NVARCHAR(10),IsPast))  
                    from Trans
                    group by AccountHeadID, TransType, IsPast
                    order by AccountHeadID, TransType, IsPast
            FOR XML PATH(''), TYPE
            ).value('.', 'NVARCHAR(MAX)') 
        ,1,1,'')

--select @cols

set @query = 'SELECT CustID,' + @cols + ' 
            from 
            (
              SELECT 
                a.CustID,
                CONVERT(NVARCHAR(10),AccountHeadID) + N''-'' + TransType + N''-'' + CONVERT(NVARCHAR(10),IsPast) Acct,
                 a.Amount
              FROM Trans a
            ) x
            pivot 
            (
                sum(Amount)
                for Acct in (' + @cols + ')
            ) p '

execute sp_executesql @query;

SQL Fiddle