年月列发出SQL

Question

我收到了一组类似的数据，如下所示。如您所见，月份在自己的列中，年份在自己的列中，然后存储产品的月度销售数据。我使用 Oracle 来获取数据。

我现在的问题是，由于销售值列在月份列下，因此很难进一步处理数据，例如在 PowerBi 中。我的目标是改变底部表格中的结构。目前我不知道它会如何工作。欢迎任何提示。

Product	Year	Jan	Feb	March	April	May	June	July	Aug	Sep	Oct	Nov	Dec
Bike	2020	12	42	42	42	20	20	12	10	15	3	16	27
Bike	2021	10	11	15	53	30	20	10	10

目标：

Product	Sale	Year	Month
Bike	12	2020	Jan
Bike	42	2020	Feb
Bike	42	2021	Jan
Bike	42	2021	Feb

Answer 1

您可以使用横向连接：

select t.product, t.year, x.month, x.sale
from t cross join lateral
     (select 'Jan' as month, t.jan as sale from dual union all
      select 'Feb' as month, t.feb as sale from dual union all
      select 'Mar' as month, t.Mar as sale from dual union all
      select 'Apr' as month, t.Apr as sale from dual union all
      select 'May' as month, t.May as sale from dual union all
      select 'Jun' as month, t.Jun as sale from dual union all
      select 'Jul' as month, t.Jul as sale from dual union all
      select 'Aug' as month, t.Aug as sale from dual union all
      select 'Sep' as month, t.Sep as sale from dual union all
      select 'Oct' as month, t.Oct as sale from dual union all
      select 'Nov' as month, t.Nov as sale from dual union all
      select 'Dec' as month, t.dec as sale from dual
     ) x;

Here 是一个 dbfiddle。

Answer 2

您可以使用 buldin 函数来取消透视列。

转到转换，选择要取消透视的列，然后使用“取消透视列”

输出：

此处为 excel 高级代码（包含您的示例数据）

let
    Source = Excel.Workbook(File.Contents("C:\Users\Mers\Downloads\pbi_tst.xlsx"), null, true),
    Sheet1_Sheet = Source{[Item="Sheet1",Kind="Sheet"]}[Data],
    #"Promoted Headers" = Table.PromoteHeaders(Sheet1_Sheet, [PromoteAllScalars=true]),
    #"Changed Type" = Table.TransformColumnTypes(#"Promoted Headers",{{"Product", type text}, {"Year", Int64.Type}, {"Jan", Int64.Type}, {"Feb", Int64.Type}, {"March", Int64.Type}, {"April", Int64.Type}, {"May", Int64.Type}, {"June", Int64.Type}, {"July", Int64.Type}, {"Aug", Int64.Type}, {"Sep", Int64.Type}, {"Oct", Int64.Type}, {"Nov", Int64.Type}, {"Dec", Int64.Type}}),
    #"Unpivoted Columns" = Table.UnpivotOtherColumns(#"Changed Type", {"Product", "Year"}, "Attribute", "Value")
in
    #"Unpivoted Columns"

这部分正在做的工作：

  #"Unpivoted Columns" = Table.UnpivotOtherColumns(#"Changed Type", {"Product", "Year"}, "Attribute", "Value")

Answer 3

这称为反透视。

测试数据：

create table sales (product, year, jan, feb, march, april, may, june, july, aug, sep, oct, nov, dec)
as
select 'Bike', 2020, 12, 42, 42, 42, 20, 20, 12, 10, 15, 3, 16, 27 from dual union all
select 'Bike', 2021, 10, 11, 15, 53, 30, 20, 10, 10, null, null, null, null from dual;

查询：

select *
from   sales s
unpivot (month
         for total in
         (jan, feb, march, april, may, june, july, aug, sep, oct, nov, dec)
        )
order by product, year;

PRODUCT       YEAR TOTAL      MONTH
------- ---------- ----- ----------
Bike          2020 JAN           12
Bike          2020 FEB           42
Bike          2020 DEC           27
Bike          2020 NOV           16
Bike          2020 OCT            3
Bike          2020 SEP           15
Bike          2020 AUG           10
Bike          2020 JULY          12
Bike          2020 JUNE          20
Bike          2020 MARCH         42
Bike          2020 APRIL         42
Bike          2020 MAY           20
Bike          2021 APRIL         53
Bike          2021 MARCH         15
Bike          2021 FEB           11
Bike          2021 JAN           10
Bike          2021 JULY          10
Bike          2021 JUNE          20
Bike          2021 MAY           30
Bike          2021 AUG           10

棘手的部分是按日历顺序对月份进行排序，因为它们的名称不一致（'Jan'、'Feb'，然后是'March'、'April'）。您可以摆弄to_date 和substr，或者使用case 表达式为每个人分配一个数字。

另外，它应该如何处理具有空值的月份？