Using ETL (non-MS) to get data from Infopath forms stored in Sharepoint 2007

Question

I'm looking at the architecture for a DW project and there will be the need for some manual collection of [structured] data eg the monthly accounting results from a country manager where they need to complete a form and fill in half a dozen values etc.

I really like the idea of using SP and InfoPath for this as it gives the security, the workflow and the customisability etc that mean it can be easily deployed as the client already has SP rolled out. The bit I am less clear on is how, technically, we might interface to the SP workflows and the forms themselves. Ideally the data would end up dropped into a database and we would use our [their!] standard ETL (DataStage, possibly sat on a linux server) via ODBC and pick it up like any other datasource but I am not sure what this requires on the SP side. The alternative would be to get at the XML of the individual forms and pull the info from there.

Are these appaoches feasible? What would need to be set up on the SP side in order to make this integration as robust and seamless as possible? Can anyone point me at docs/reading matter that might give me some more background info?

Thanks, Dex

Answer 1

First up, accessing sharepoint's databases is never the answer to any integration question. You should treat it as a black box.

So, how should you get the data? Web Services + HTTP. SharePoint offers a large amount of Web services to get at the data you need. If you're working with IP forms, then ultimately you will need to grab the resultant XML file from the document library and parse it to get the data you need. The Web services can be used to enumerate the IP forms, and you can use straight HTTP to grab to xml file. This is probably the approach that would be offered by most experienced sharpepoint people.