TM1 - methodology for publishing cubes

[dssd]

I have a requirement where my existing TM1 cubes are going agaisnt desparate data sources
sucha as text files, csvs, excel files, etc. We are consolidating all these sources
to a datawarehouse. I was wondering how easy or hard it could be to just change the
mapping in TM1 to go from the files to the datawarehouse elements

Second query, what is the methodology for using TM1. In case of transformer we used to import objects in
Framework Manager, then create reports which would act as data sources in Cubes, then
publish the cube as a package from FM. Is it true for TM1 as well or should we directly publish cubes as packages from TM1?

[dusherwo]

Not a big deal - set up an ODBC DSN (64 bit if your TM1 server is 64 bit) for the warehouse, create copies of your text-based TIs replace the data source by ODBC, write the query, test and fix.
As far as the 'methodology' for using TM1, have you decided how you are going to deliver output to users? If you use Excel then you are done. If you want to deliver through BI then you define a datasource for each cube you work with. It's not my area of expertise, but I don't think you are obliged to use FM.
Watch out too - the MDX interface between BI pre 10.1 and TM1 pre 9.5.2 wasn't very good when pushed eg suppressing zeros on the query output - yuck.

[dssd]

Whats a TI in "text based TIs"

[OLAPBPMguy]

TI = "Turbo Integrator" a.k.a TM1's built in ETL module.

The thing you will have to remap in other words to convert from flat file input to ODBC input to the cubes ....

[bloggerman]

Is it a good practice to source TM1 from a dimensional datawarehouse as compared to flat files

[dusherwo]

@dssd....
If you don't know what TI stands for, _how_ are you loading your data (and structures!) now? DBSW? Process Worksheets!??
@bloggerman, I would suggest keeping the number of stages between the data and TM1 down to the minimum. If your data comes to you as flat files, use them, unless you have lots of repair work to do to them. TI is not a good tool for data repair, in my view. Also flat files will read in much faster than ODBC.