3

I am working for a project using the BI tools of Pentaho: Pentaho Mondrian.

I am experiencing problems with the use of aggregate tables: for make aggregate tables I used Pentaho Aggregate Designer and I successfully created the DDL, populated the new tables and published the schema into Pentaho.

I am using Saiku but I'd see, for the response times, that it doesn't use the aggregate tables even if it's specified in the schema.

Of course I changed the values in every mondrian.properties file,

mondrian.rolap.aggregates.Use=true
mondrian.rolap.aggregates.Read=true

but no way for make it working!

I attached you a screenshot of the aggregate designer and another of the result using Pentaho and Saiku (you can see that the query was executed in more than 10s when the aggregate table has only 9 tuples)

Do you have some ideas or did you experienced this problem and you solved?

Thank you in advance!

enter image description here

enter image description here

flatlopa
  • 111
  • 10

1 Answers1

1

It took me also a time to make aggregate tables work. Here is one quick tip if you face this issue:

  • Use standalone version of Saiku server.
  • Deploy your OLAP schema on Saiku.
  • Enable Special Log File specifically for Mondrian SQL Statements in saiku/tomcat/webapps/saiku/WEB-INF/clasess/log4j.xml
  • Run Saiku, execute same queries where you expect the aggregate tables will be used.
  • Check the log saiku/tomcat/logs/mondrian.sql to see if your aggregate tables are really used or not.

This helped me a lot when finding the solution for my problem in OLAP schema (wrong db column used for aggregate dimension).

mzy
  • 1,754
  • 2
  • 20
  • 36