Question

我有一个针对我的数据源的查询，该数据源返回一个维度（例如，团队名称），以及与相同维度的十字架相关联的若干度量（例如，胜利，亏损）。例如：

+===========+===============+===============+===============+===============+===========+===========+
|           | Cardinals - W | Cardinals - L | Blue Jays - W | Blue Jays - L | Hawks - W | Hawks - L |
+===========+===============+===============+===============+===============+===========+===========+
| Cardinals | x             | x             | 5             | 10            | 1         | 2         |
+-----------+---------------+---------------+---------------+---------------+-----------+-----------+
| Blue Jays | 10            | 5             | x             | x             | 8         | 4         |
+-----------+---------------+---------------+---------------+---------------+-----------+-----------+
| Hawks     | 2             | 1             | 4             | 8             | x         | x         |
+===========+===============+===============+===============+===============+===========+===========+

我想要做的就是让这个不透明，给我'A'和'B'列以及非团队特定的'W'和'L'列。例如：

+===========+===========+====+====+
| Team A    | Team B    | W  | L  |
+===========+===========+====+====+
| Blue Jays | Cardinals | 10 | 5  |
+-----------+-----------+----+----+
| Blue Jays | Hawks     | 8  | 4  |
+-----------+-----------+----+----+
| Cardinals | Blue Jays | 5  | 10 |
+-----------+-----------+----+----+
| Cardinals | Hawks     | 1  | 2  |
+-----------+-----------+----+----+
| Hawks     | Blue Jays | 4  | 8  |
+-----------+-----------+----+----+
| Hawks     | Cardinals | 2  | 1  |
+===========+===========+====+====+

（因为我的用例实际上有点不同，我实际上也喜欢{Team A=Blue Jays, Team B=Blue Jays, W=x, L=x}的行。

如果我在查询编辑器中执行Unpivot，我最终会获得

+===========+===============+=======+
|           | Attribute     | Value |
+===========+===============+=======+
| Cardinals | Cardinals - W | x     |
+-----------+---------------+-------+
| Cardinals | Cardinals - L | x     |
+-----------+---------------+-------+
| Cardinals | Blue Jays - W | 5     |
+-----------+---------------+-------+
| Cardinals | Blue Jays - L | 10    |
+-----------+---------------+-------+
| Cardinals | Hawks - W     | 1     |
+-----------+---------------+-------+
| Cardinals | Hawks - L     | 2     |
+-----------+---------------+-------+
| Blue Jays | Cardinals - W | 10    |
+-----------+---------------+-------+
| Blue Jays | Cardinals - L | 5     |
+-----------+---------------+-------+
| Blue Jays | Blue Jays - W | x     |
+-----------+---------------+-------+
| (more rows removed for brevity)   |
+===========+===============+=======+

有没有合理的方法在我的数据集中以这种方式实现数据透视？

Answer 1

与Marc的答案类似，但步数较少。

从不透明的数据开始。

选择属性列，然后选择拆分列 - ＆gt;通过Delimiter。 PowerBI通常足够智能自动检测分隔符，但如有必要，您可以在对话框中对其进行编辑。

然后选择Attribute.2列并选择Pivot Column。将值列更改为值，并在高级选项下，将聚合值函数更改为不聚合。

然后你去。

Answer 2

要做到这一点：

我在查询编辑器中以Table1开头：

然后我选择了Column1和Unpivoted Other Columns：

...产生了这个：

然后我添加了两个新列：一个是从属性列中提取的W或L，另一个是仅从属性列中提取的团队名称。我使用Text.End和Text.Start进行了这些提取：

然后我选择了自定义列并对其进行了旋转，并为关联值选择了“值”列：

...产生了这个：

然后我删除了属性列。然后我选择了W列并填满了。然后我选择了L列和Filled Down。然后我选择了W列并过滤掉了x。然后我右键单击表的左上角并选择Remove Duplicates。最后，我将Column1和Custom.1列重命名为A队和B队。

这是我的查询代码：

let
Source = Excel.CurrentWorkbook(){[Name="Table1"]}[Content],
#"Changed Type" = Table.TransformColumnTypes(Source,{{"Column1", type text}, {"Cardinals - W", type any}, {"Cardinals - L", type any}, {"Blue Jays - W", type any}, {"Blue Jays - L", type any}, {"Hawks - W", type any}, {"Hawks - L", type any}}),
#"Unpivoted Other Columns" = Table.UnpivotOtherColumns(#"Changed Type", {"Column1"}, "Attribute", "Value"),
#"Added Custom" = Table.AddColumn(#"Unpivoted Other Columns", "Custom", each Text.End([Attribute],1)),
#"Added Custom1" = Table.AddColumn(#"Added Custom", "Custom.1", each Text.Start([Attribute], Text.Length([Attribute])-4)),
#"Pivoted Column" = Table.Pivot(#"Added Custom1", List.Distinct(#"Added Custom1"[Custom]), "Custom", "Value"),
#"Removed Columns" = Table.RemoveColumns(#"Pivoted Column",{"Attribute"}),
#"Filled Up" = Table.FillUp(#"Removed Columns",{"W"}),
#"Filled Down" = Table.FillDown(#"Filled Up",{"L"}),
#"Filtered Rows" = Table.SelectRows(#"Filled Down", each ([W] <> "x")),
#"Removed Duplicates" = Table.Distinct(#"Filtered Rows"),
#"Renamed Columns" = Table.RenameColumns(#"Removed Duplicates",{{"Column1", "Team A"}, {"Custom.1", "Team B"}})
in
#"Renamed Columns"

我希望这会有所帮助。

取消激活Power BI表以将列减少到仅测量

2 个答案: