没有聚合的 SQL Server Pivot Data答案

【问题标题】：SQL Server Pivot Data without aggregation没有聚合的 SQL Server Pivot Data
【发布时间】：2019-07-03 17:24:54
【问题描述】：

我知道这已经被问过很多次了，但我很难找到我需要的解决方案。

数据：

Id   FKId   TypeId   Score   ScoreDate
1    317    1        90      2019-01-01
2    317    1        80      2019-01-02
3    317    2        65      2019-01-03
4    317    2        80      2019-01-04
5    318    1        82      2019-01-01
6    318    1        78      2019-01-02
7    318    2        55      2019-01-03
8    318    2        63      2019-01-04

预期结果：

FkId   A1   A2   B1  B2
317    90   80   65  80
318    82   78   55  63

我尝试使用枢轴来获取数据，但我偶然发现的是您需要使用聚合函数，这会导致错误的结果。其中一种尝试的示例：

select *
from 
(
  select [FkId], [TypeId], [Score]
  from #yt
) src
pivot
(
  max([Score])
  for [TypeId] in ([1], [2], [3])
) piv;

query result:
FkId  1     2       3
101   83    82  NULL
102   96    80  NULL

我知道我需要不止一个查询才能获得预期结果，但这不是我关心的问题，现在我只想将一种类型的分数与每个分数转置到它自己的列中。

任何帮助都可以。

【问题讨论】：

标签： sql-server tsql dynamic pivot

【解决方案1】：

试试这个：

CREATE TABLE #yt
(
    [Id] INT
   ,[FKId] INT
   ,[TypeId] TINYINT
   ,[Score] INT
   ,[ScoreDate] DATE
)

INSERT INTO #yt
VALUES (1, 317, 1, 90, '2019-01-01')
      ,(2, 317, 1, 80, '2019-01-02')
      ,(3, 317, 2, 65, '2019-01-03')
      ,(4, 317, 2, 80, '2019-01-04')
      ,(5, 318, 1, 82, '2019-01-01')
      ,(6, 318, 1, 78, '2019-01-02')
      ,(7, 318, 2, 55, '2019-01-03')
      ,(8, 318, 2, 63, '2019-01-04');

select *
from 
(
    SELECT [FKId]
        ,CHAR(64 + DENSE_RANK() OVER (PARTITION BY FKID ORDER BY TypeID)) + CAST(DENSE_RANK() OVER (PARTITION BY FKID, TypeID ORDER BY [ScoreDate]) AS VARCHAR(12))
        ,[Score]
    FROM #yt
) src ([FKId], [TypeID], [Score])
pivot
(
  max([Score])
  for [TypeId] in ([A1], [A2], [B1], [B2])
) piv;


DROP TABLE #yt;

或者如果你想让它动态化：

DECLARE @DynamicTSQLStatement NVARCHAR(MAX)
       ,@DynamicColumns NVARCHAR(MAX)

SET @DynamicColumns = STUFF
                        (
                            (
                                SELECT DISTINCT ',[' + CHAR(64 + DENSE_RANK() OVER (PARTITION BY FKID ORDER BY TypeID)) + CAST(DENSE_RANK() OVER (PARTITION BY FKID, TypeID ORDER BY [ScoreDate]) AS VARCHAR(12)) + ']'
                                FROM #yt                               
                                FOR XML PATH('') ,TYPE
                            ).value('.', 'NVARCHAR(MAX)')
                            ,1
                            ,1
                            ,''
                        );

SET @DynamicTSQLStatement = N'
select *
from 
(
    SELECT [FKId]
        ,CHAR(64 + DENSE_RANK() OVER (PARTITION BY FKID ORDER BY TypeID)) + CAST(DENSE_RANK() OVER (PARTITION BY FKID, TypeID ORDER BY [ScoreDate]) AS VARCHAR(12))
        ,[Score]
    FROM #yt
) src ([FKId], [TypeID], [Score])
pivot
(
  max([Score])
  for [TypeId] in ('+@DynamicColumns+')
) piv;'


EXEC sp_executesql @DynamicTSQLStatement;

【讨论】：

【解决方案2】：

检查一下：我使用了两个查询，然后将它们合并并取其最大值

CREATE TABLE t (
  "Id" INTEGER,
  "FKId" INTEGER,
  "TypeId" INTEGER,
  "Score" INTEGER,
  "ScoreDate" date
);


GO

✓

INSERT INTO t
  ("Id", "FKId", "TypeId", "Score", "ScoreDate")
VALUES
  ('1', '317', '1', '90', '2019-01-01'),
  ('2', '317', '1', '80', '2019-01-02'),
  ('3', '317', '2', '65', '2019-01-03'),
  ('4', '317', '2', '80', '2019-01-04'),
  ('5', '318', '1', '82', '2019-01-01'),
  ('6', '318', '1', '78', '2019-01-02'),
  ('7', '318', '2', '55', '2019-01-03'),
  ('8', '318', '2', '63', '2019-01-04');
GO

受影响的 8 行

select [FkId], max([1]) as c1, max([2]) as c2, max([3]) as c3, max([4]) as c4 from
(
select FkId, [1], [2], [3], [4] from
(
  select *
  from
  (
    select [FkId], [TypeId], [Score]
    from t
  ) src
  pivot
  (
    max([Score])
    for [TypeId] in ([1], [2] , [3], [4])
  ) as Piv1
) P
union all
select FkId, null, null, [1], [2] from
(
  select *
  from
  (
    select [FkId], [TypeId], [Score]
    from t
  ) src
  pivot
  (
    min([Score])
    for [TypeId] in ( [1], [2])
  ) as Piv2
) P2
) PP
group by FkId
GO

FkId | c1 | c2 | c3 | c4 ---: | -: | -: | -: | -： 317 | 90 | 80 | 80 | 65 318 | 82 | 63 | 78 | 55

db小提琴here

【讨论】：