Question

在将RANKX函数与模型中特定表的列组合时，我遇到了笛卡尔积问题。

我使用Power BI Desktop的2.53.4954.621版本连接到内部部署数据仓库。我的模型中有四个表，Sales，Region，Sales Goal和Calendar。

所有表都加入了Sales，因此我将为该实体提供相关的表结构。

Sales:
rowid
date
rep_id
region_id
sales
discount
company number
RepRegionKey (merging of rep_id and region_id)

以下是我的表格之间定义的关系

1. From Sales (date) to Calendar (Date), *:1, Single direction
2. From Sales (region_id) to Region (id), *:1, Both directions
3. From Sales (RepRegionKey) to Sales Goal (RepRegionKey), *:1, Both directions

我也在Sales表上定义了一些度量。我不确定哪些可能很重要，所以我会将它们全部列出来。

年初至今销售额 = CALCULATE(SUM('Sales'[sales]) - SUM('Sales'[discount]))

YE销售目标 =

CALCULATE(SUM('Sales Goal'[sales_goal]))

$ YTD目标与年初至今的销售额 =

CALCULATE([YTD Sales] - [YE Sales Goal])

年初至今的目标与年初至今的销售额 = DIVIDE([$ YTD Goal vs YTD Sales],[YE Sales Goal])

差异等级 = RANKX(ALL('Sales Goal'[rep_id]),CALCULATE([% YTD Goal vs YTD Sales]),,,DENSE)

当我将 Diff Rank 度量添加到包含＆＃39; Region＆＃39; [region_desc]（或任何＆＃39; Region＆＃39;除了company_number之外的字段，表中只有一个值。

在下面的图片中观察region_desc字段如何与％YTD目标与YTD销售度量和平共享一个表（图中最左边的表），在该度量上构建RANX功能。此外，在中间表中，您会看到RANKX函数在没有region_desc字段的情况下正常工作。但是，在右侧的最终表中，组合RANKX和region_desc字段时会出现笛卡尔积。

RANKX Cartesian Product

我希望看到每个Rep的准确排名，如下：

Desired Output

我的关系是否不正确，或者我是否需要调整RANKX定义？

这些是我尝试过的定义，有不同程度的失败：

---没有REGION_DESC的工作

Diff Rank = RANKX(ALL('Sales Goal'[rep_id]),CALCULATE([% YTD Goal vs YTD Sales],ALLSELECTED('Region'[region_desc])),,,DENSE) 

Diff Rank = RANKX(ALL('Sales Goal'[rep_id]),CALCULATE([% YTD Goal vs YTD Sales],ALL(Region[region_desc])),,,DENSE)

Diff Rank = RANKX(ALL('Sales Goal'[rep_id]),CALCULATE([% YTD Goal vs YTD Sales],FILTER(ALL(Region[region_desc]),[region_desc]<>"BUBBA")),,,DENSE)

Diff Rank = RANKX(ALL('Sales Goal'[rep_id]),CALCULATE([% YTD Goal vs YTD Sales],ALLEXCEPT('Region','Region'[region_desc])),,,DENSE)

---不使用REGION_DESC

Diff Rank = RANKX('Sales Goal',CALCULATE([% YTD Goal vs YTD Sales],ALLEXCEPT('Region','Region'[region_desc])),,,DENSE)

Diff Rank = RANKX('Sales Goal',CALCULATE([% YTD Goal vs YTD Sales],ALLSELECTED('Region'[region_desc])),,,DENSE)

Diff Rank = RANKX('Sales Goal',CALCULATE([% YTD Goal vs YTD Sales],ALL('Region'[region_desc])),,,DENSE)

Diff Rank = RANKX('Sales Goal',CALCULATE([% YTD Goal vs YTD Sales],FILTER(ALL(Region[region_desc]),[region_desc]<>"BUBBA")),,,DENSE)

Diff Rank = RANKX(DISTINCT('Sales Goal'[rep_id]),CALCULATE([% YTD Goal vs YTD Sales]),,,DENSE)

Diff Rank = RANKX(VALUES('Sales Goal'[rep_id]),CALCULATE([% YTD Goal vs YTD Sales]),,,DENSE)

提前谢谢！

RANKX函数在Power BI中生成笛卡尔积

0 个答案: