PYJ file size

[nayitian]

Hi

I noticed that every time after cube being generated by invoking .pyj file, the file size is dramatically increased from 60K-ish to 100M.
How can I set up to keep the pyj file size as minimum as it was created?

Thank you very much.

[Arsenal]

the increased size is due to a possibly very large number categories for your cube.
One way to keep the size down would be to drop the save option from your cube build command line so the categories being generated are not saved, but this may lead to increased cube build times since the categories will have to be generated every time you build the cube.

You may also try out using an MDL instead of .pyj (less chances of mdl getting corrupted and mdl will use less space for the same model as .pyj), continuing to save the categories and then perdiodically using the clean house option (be careful since manually created categories, alternate drill paths etc, can get wiped out too) and/or manually deleting the categories periodically

[nayitian]

Thank you so much for clarifying on this.

A little more help I am asking for here,
When I generate cube by using .pyj, the parameter I use is to define log file location. Is there any parameter that can be used to clean generated categories after cube being created?

[Arsenal]

Thank you so much for clarifying on this.

A little more help I am asking for here,
When I generate cube by using .pyj, the parameter I use is to define log file location. Is there any parameter that can be used to clean generated categories after cube being created?

Look at the following thread. The OP's issue is something else, but he too is asking about deleting categories after every cube generation.

http://www.cognoise.com/community/index.php/topic,16786.0.html

[Arsenal]

Thank you so much for clarifying on this.

A little more help I am asking for here,
When I generate cube by using .pyj, the parameter I use is to define log file location. Is there any parameter that can be used to clean generated categories after cube being created?

btw, are you not using the save model file parameter at all during cube build?
If not, then all you have to do is delete all the categories in your model once manually, upload it to build location and then use this new model for subsequent cube builds. Due to the absence of the save model parameter, the cube will be forced to generate categories during every cube build but not save these categories

[nayitian]

Most of my cubes are daily refreshed, the categories can vary from day one to day two, that's the reason I let the categories re-created every time it runs.
Thanks a lot.

[AussiePete2011]

Have a read of this document

Transformer in Production
http://www.ibm.com/developerworks/data/library/cognos/page354.html

Cheers
Peter