Digging Mondrian

On Friday I committed code to 3.0 trunk to allow people to execute an MDX query on a Mondrian server and get the result back in a tabular format. This particular code to “flatten” an OLAP cube into rows was written by Julian Hyde, the lead developer and founder of Mondrian OLAP a.k.a. Pentaho analyses.

If you run the Pentaho demo on your box and then look at the Analyses sample, you could see something like this:

JPivot sample

Suppose you wanted to get this exact data to work with, create analytical data, exports, … Well, now you have the option of doing it in Kettle:

Mondrian input

What you do is create a database connection to the database that Mondrian reads from, hand it the location of the Mondrian schema (catalog). You then click on the MDX button in JPivot and copy the MDX query to the “Mondrian Input” step in Kettle. That’s all it takes.

You then can preview the output of the Dummy step, like this in screen shot:

Mondrian input step preview results

As you can see, integration of the different Pentaho software pieces is continuing…

Until next time!

Matt

2 comments

  • lol!

    Another example of the “If i tell you i will have to kill you” technologicals advances of Kettle. Is there no limit to what you can do?

    Anyway, looking forward to the 3.0 release!!!

    All the best

  • Nice Matt !

    And more than nice : usefull, especially if you can support a MDX code generation interface.
    2007 is definitly the “year of integration” between the different Pentaho packages.

    Next piece to be integrated should be Pentaho Metadata files, already available in different software such as Pentaho Report Wizard 1.6 or FreeAnalysis Schema Designer,

    Regards,
    Patrick