Hi Mandar,
We're essentially saying you can do it in several ways, each having pro's and con's.
Please consider these options:
With PCM rules:
- manageable by PCM model builders (with a little PCM Rule/vbscript knowledge)
- depending on your PCM model requirements, might need a seperate PCM gridvalue, PCM versions or a seperate PCM 'loaded' model
- the split/ratio can made be variable, based on other PCM data, allowing different cost model simulations
- in high volume of data, this means many different rules, there might be performance concerns
With FIM:
- mapping is easy to maintain by business users
- FIM is very performant, using best-practice ETL engine (BODS) and DataLoader underneath
- only end result is used in the PCM model, no model performance concerns in high volumes
- provides drill through feature (any loaded cell in PCM can be traced back to a report on ECC)
With BODS:
- could in theory do the same job as FIM, but the job itself is harder to develop (connection to PCM) this because you need to know the PCM web services to call (Dataloader) and drill through tables etc.
- cannot be maintained by business users
- could possibly use any other source of data do get the ratio/split
p.s. best-practice recommendation is still to use FIM to create this job, and if really needed (if te FIM mapping is not sufficient), you can still customize it afterwards with BODS.
With Databridge SPE-files:
- The ratio/split would need to be fixed in the SPE file
- I haven't got an exact example of how to do this, but it proves not to be straightforward
- performance is questionable with high volumes, since DataLoader routines are not used in the background. You might need a Bulk table load, which you have to be carefull with (less consistency checks)