我有一个正常的选择查询，结果如下输出。
select cid,x1,x2,x3,x4,fy
from temp_table;
cid x1 x2 x3 x4 fy
----------------------------
6657 100 0 0 200 2014
6658 300 0 0 400 2015
6659 500 0 0 600 2016
我希望它重写它打印以下输出。
2014 2015 2016
-------------------------
x1 100 300 500
x2 0 0 0
x3 0 0 0
x4 200 400 600
如何做到这一点？
【问题讨论】：
这是一种只使用子查询和聚合的方法：
select name,
sum(case when fy = 2014 then x end) as "2014",
sum(case when fy = 2015 then x end) as "2015",
sum(case when fy = 2016 then x end) as "2016"
from (select fy,
(case when n.n = 1 then 'x1'
when n.n = 2 then 'x2'
when n.n = 3 then 'x3'
when n.n = 4 then 'x4'
end) as name,
(case when n.n = 1 then x1
when n.n = 2 then x2
when n.n = 3 then x3
when n.n = 4 then x4
end) as x
from temp_table cross join
(select 1 as n from dual union all
select 2 from dual union all
select 3 from dual union all
select 4 from dual
) n
) t
group by name;
您也可以使用pivot，但这是最近才添加到 Oracle SQL 中的，所以我倾向于使用这种方法。
【讨论】：
pivot 来做这件事我正在使用oracle 11g。可能存在我们的查询输出中有 n 年的情况，因此转置查询应该有 n+1 列
更通用的解决方案是UNPIVOT and PIVOT。
您的 temp_table 已经是一个以 (x1 x2 x3 x4) 作为 x-axis 和 fy 作为 y-axis 的数据透视表。
首先我们需要将temp_table UNPIVOT 到unpivoted_temp_table，然后以 fy 作为 x 轴和 (x1 x2 x3 x4) 作为 y 轴进行 PIVOT： p>
with unpivoted_temp_table as (
SELECT *
FROM temp_table
UNPIVOT (
totalSales
FOR x
IN (x1, x2, x3, x4)
)
)
select *
FROM unpivoted_temp_table
PIVOT (
SUM(totalSales)
FOR fy
IN (2014, 2015, 2016)
)
order by 1 --order by column X
【讨论】：
在 Oracle 11G 中转置的三个可能选项
解码选项
一个。如果要转置，请对新列使用 SUM 和 DECODE 并总结
b.如果要转置，请对新列使用 MAX 和 DECODE 而不是总结
使用 PIVOT
如 Gordon 所说，使用 WITH 和 SUB SELECT
要使用数据透视，在您的代码中是不可能的，因为您的初始数据集已经被透视。
WITH TEMP1
AS (SELECT
6657 AS CID,
100 AS VAL,
'X1' AS XCORD,
2014 AS FY
FROM
DUAL
UNION ALL
SELECT
6657 AS CID,
0 AS VAL,
'X2' AS XCORD,
2014 AS FY
FROM
DUAL
UNION ALL
SELECT
6657 AS CID,
0 AS VAL,
'X3' AS XCORD,
2014 AS FY
FROM
DUAL
UNION ALL
SELECT
6657 AS CID,
200 AS VAL,
'X4' AS XCORD,
2014 AS FY
FROM
DUAL
UNION ALL
SELECT
6658 AS CID,
300 AS VAL,
'X1' AS XCORD,
2015 AS FY
FROM
DUAL
UNION ALL
SELECT
6658 AS CID,
0 AS VAL,
'X2' AS XCORD,
2015 AS FY
FROM
DUAL
UNION ALL
SELECT
6658 AS CID,
0 AS VAL,
'X3' AS XCORD,
2015 AS FY
FROM
DUAL
UNION ALL
SELECT
6658 AS CID,
400 AS VAL,
'X4' AS XCORD,
2015 AS FY
FROM
DUAL
UNION ALL
SELECT
6659 AS CID,
500 AS VAL,
'X1' AS XCORD,
2016 AS FY
FROM
DUAL
UNION ALL
SELECT
6659 AS CID,
0 AS VAL,
'X2' AS XCORD,
2016 AS FY
FROM
DUAL
UNION ALL
SELECT
6659 AS CID,
0 AS VAL,
'X3' AS XCORD,
2016 AS FY
FROM
DUAL
UNION ALL
SELECT
6659 AS CID,
600 AS VAL,
'X4' AS XCORD,
2016 AS FY
FROM
DUAL)
SELECT
*
FROM
(SELECT
XCORD,
FY,
SUM ( VAL ) AS VAL
FROM
TEMP1
GROUP BY
XCORD,
FY) PIVOT (SUM ( VAL ) FOR FY IN ('2014', '2015', '2016'))
ORDER BY
XCORD;
结果：
XCORD '2014' '2015' '2016'
----- ------ ------ ------
X1 100 300 500
X2 0 0 0
X3 0 0 0
X4 200 400 600
如果你查看数据集，你可以看到结果是
CID VAL XCORD FY
---- ---- ----- ----
6657 100 X1 2014
6657 0 X2 2014
6657 0 X3 2014
6657 200 X4 2014
6658 300 X1 2015
6658 0 X2 2015
6658 0 X3 2015
6658 400 X4 2015
6659 500 X1 2016
6659 0 X2 2016
6659 0 X3 2016
6659 600 X4 2016
这或多或少是您数据的 UNPIVOT。
【讨论】：
您可以结合使用 unpivot 和 pivot 来做到这一点。
select * from
(select * from
(select x1,x2,x3,x4,fy from table1)
unpivot(val for x in (x1,x2,x3,x4)))
pivot(sum(val) for fy in (2014,2015,2016))
【讨论】：