I'm at a client site where TMM 2.01.05.32 will be evaluated. Currently, the data mapping process is comprised of a single excel spreadsheet that accommodates 3 versions of a given table. From that spreadsheet (1 per source data spec), three different ERwin models are created. File (inbound as-is), Stage (transformation layer), and Base (final target DB).
What's Teradata best practice for setting up TMM given the above scenario?
1 project per inbound spreadsheet with 3 subject areas (file, stage & base) to accommodate the three entities/tables created?
1 project per per each unique table structure? i.e., 3 projects per table
FYI - inbound spec is the driver for the File table, File table drives the Stage table and Stage drives the Base table structure. The spreadsheet defines the source attribute names, and what attributes should flow to each table version.
Any thoughts as to the best way to setUP TMM to work with the above?
Thanks for any insight or help offered.
You can only map one object to another if they are both in the same project, so you would typically create separate data representations (DRs) for File, Stage and Base, then map File to Stage and map Stage to Base. You can import the DRs using the blank templates provided by TMM installer or you can use pre-defined ERwin reports to generate the import files if you are using ERwin.
If you have 3 models for each source, then you would have an additional 3 DRs for each source.
I recommend the TMM tutorial to teach you the basics - it is found in pdf format under My Documents\Teradata Mapping Manager\Documents.