[Mondrian] MDX performance - best practises

Kleyson de Sousa Rios krios at pentaho.com
Fri Sep 30 06:15:23 EDT 2016


Regarding Mondrian 3.x, are there some set of function that could give us a better performance if used instead of another ones ?

For example, the following MDX’s give me the same result:

WITH

-- Rows Sets
SET [_SELECTED_ROWS_SET] as  [People].[Name].Members
MEMBER [Measures].[Value] AS FILTER(
                                                             [_SELECTED_ROWS_SET]
                                                             ,  [Measures].[Total Time] > 10
                                               ).Count
-- Columns Sets
SET [COLUMNS_SET] as { [Measures].[Value] }
SELECT
  [COLUMNS_SET] ON COLUMNS
FROM
  [CUBE]

and

WITH

-- Rows Sets
SET [_SELECTED_ROWS_SET] as  [People].[Name].Members
MEMBER [Measures].[Value] AS SUM(
                                                          [_SELECTED_ROWS_SET]
                                                          ,IIF( [Measures].[Total Time] > 10, 1, null)
                                               )

-- Columns Sets
SET [COLUMNS_SET] as { [Measures].[Value] }
SELECT
  [COLUMNS_SET] ON COLUMNS
FROM
  [CUBE]



I mean, using SUM+IIF should be better than using FILTER+COUNT ?

I’m asking this because some documentations about performance in SSAS MDX say that some functions give a better performance than others, such as, SUM+IIF instead of FILTER+COUNT.

I’m wondering to know if we should care about this concept when writing MDX Mondrian queries.

Best Regards.

Kleyson Rios

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.pentaho.org/pipermail/mondrian/attachments/20160930/0a093090/attachment.html 


More information about the Mondrian mailing list