Question

我正在学习MDX，我有一个问题要创建一个百分比。基本上，我从Google Analytics API中提取多组网页的数据。对于每个集合，我提取UniquePageviews的数量。

SQL列看起来像这样：

日期
deviceCategory
介质
uniquePageviews
交易
set_of_page（行包含页面集的名称。简化：1。首先，2。第二，3。）
品牌（品牌A，品牌B）
市场（市场X，市场Y）

以下是我目前的观点： http://i.stack.imgur.com/ZTC30.png

我想要的是第1页到第2页以及第2页到第3页的UniquePageviews的百分比所以：UniquePageviews 2 / Uniquepageviews 1＆amp; UniquePageviews 3 / Uniquepageviews 2。

http://i.stack.imgur.com/xf2gc.png

您是否知道在MDX中使用这种方法的简单方法？

以下是我现在用于显示综合浏览量的代码：

WITH
SET [~COLUMNS_Brand_Brand] AS
SET [~COLUMNS_Market_Market] AS
    {[Market].[Market].Members}
SET [~COLUMNS_set_of_page segment_set_of_page] AS
    {[set_of_page].[set_of_page].Members}
SET [~ROWS] AS
    {[Date].[Date].Members}
SELECT
NON EMPTY NonEmptyCrossJoin([~COLUMNS_Brand_Brand], NonEmptyCrossJoin([~COLUMNS_Market_Market], [~COLUMNS_set_of_page_set_of_page])) ON COLUMNS,
NON EMPTY [~ROWS] ON ROWS
FROM [MySQL_Cube]

非常感谢！

Answer 1

当你的mdx隐含地采取措施时？所有mdx查询都需要测量一些内容，因此您当前的脚本必须采用默认的多维数据集度量 - 我将其称为[Measures].[NumTransactions]。你的第一个SET似乎也是空白的。因此，我们假设完整的显式脚本是：

WITH
SET [~COLUMNS_Brand_Brand] AS {[Brand].[Brand].[All]}
SET [~COLUMNS_Market_Market] AS
    {[Market].[Market].Members}
SET [~COLUMNS_set_of_page segment_set_of_page] AS
    {[set_of_page].[set_of_page].Members}
SET [~ROWS] AS
    {[Date].[Date].Members}
SELECT
NON EMPTY NonEmptyCrossJoin([~COLUMNS_Brand_Brand], NonEmptyCrossJoin([~COLUMNS_Market_Market], [~COLUMNS_set_of_page_set_of_page])) ON COLUMNS,
NON EMPTY [~ROWS] ON ROWS
FROM [MySQL_Cube]
WHERE [Measures].[NumTransactions];

我使用的是MS SSAS，因此有点不确定是否适用于你，但只是为了帮助我，我将切换到*运算符进行交叉连接：

WITH
SET [~COLUMNS_Brand_Brand] AS {[Brand].[Brand].[All]}
SET [~COLUMNS_Market_Market] AS
    {[Market].[Market].Members}
SET [~COLUMNS_set_of_page segment_set_of_page] AS
    {[set_of_page].[set_of_page].Members}
SET [~ROWS] AS
    {[Date].[Date].Members}
SELECT
NON EMPTY 
   [~COLUMNS_Brand_Brand]
  *[~COLUMNS_Market_Market]
  *[~COLUMNS_set_of_page_set_of_page] ON COLUMNS,
NON EMPTY [~ROWS] ON ROWS
FROM [MySQL_Cube]
WHERE [Measures].[NumTransactions];

我希望（由于未经测试，我有点不确定所有上下文将被保留），我们可以添加一个自定义的meassure，将当前页面除以前一页以获得百分比：

WITH
SET [~COLUMNS_Brand_Brand] AS {[Brand].[Brand].[All]}
SET [~COLUMNS_Market_Market] AS
    {[Market].[Market].Members}
SET [~COLUMNS_set_of_page segment_set_of_page] AS
    {[set_of_page].[set_of_page].Members}
SET [~ROWS] AS
    {[Date].[Date].Members}
MEMBER [Measures].[NumTransactionsPercent] AS
     [set_of_page].CURRENTMEMBER
   / [set_of_page].CURRENTMEMBER.LAG(1)
SELECT
NON EMPTY 
   [~COLUMNS_Brand_Brand]
  *[~COLUMNS_Market_Market]
  *[~COLUMNS_set_of_page_set_of_page] ON COLUMNS,
NON EMPTY [~ROWS] ON ROWS
FROM [MySQL_Cube]
WHERE [Measures].[NumTransactionsPercent];  //[Measures].[NumTransactions];

MDX：根据同一维度

1 个答案: