Power BI矩阵视觉中DAX度量未遵循筛选上下文问题解决

场景与问题 #

我在Power BI中有一张存储过去3个月各账户周度销售额的表，包含字段：

• account：账户ID
• sales_amount：销售额
• week_started_at：每周起始日期
  已添加WeekRank列用于对每周进行排名。需要计算各账户销售额的周环比(W/W Change)，但使用的DAX度量在表格视觉中正常，在矩阵视觉中却未遵循筛选上下文，导致所有账户的周环比数值一致，不符合预期。

示例表结构 #

原问题DAX度量 #

current_week_sales = COALESCE(
    CALCULATE(SUM(table[sales_amount]), FILTER(ALLSELECTED(table), table[WeekRank] = MAX(table[WeekRank]))),0)

last_week_sales = COALESCE(
    CALCULATE(SUM(table[sales_amount]), FILTER(ALLSELECTED(table), table[WeekRank] = MAX(table[WeekRank]) -1)
    ), 0)

W/W_Change = 
VAR weekly_comparison = [current_week_sales] - [last_week_sales]
RETURN 
COALESCE(
    IF(ISBLANK(weekly_comparison), "", DIVIDE(weekly_comparison , [last_week_sales]))
    ,0)

解决方案 #

修改后的DAX度量如下，核心是将ALLSELECTED(table)改为仅清除指定列（WeekRank或week_started_at）的筛选上下文，保留账户等其他维度的筛选，确保矩阵视觉中每个账户能独立计算周环比：

current_week_sales = COALESCE(
    CALCULATE(SUM(table[sales_amount]), FILTER(ALLSELECTED(table[WeekRank]), table[WeekRank] = MAX(table[WeekRank]))),0)

last_week_sales = COALESCE(
    CALCULATE(SUM(table[sales_amount]), FILTER(ALLSELECTED(table[week_started_at]), table[WeekRank] = MAX(table[WeekRank]) -1)
    ), 0)

内容的提问来源于stack exchange，提问作者SusanD