When Normalize version 2.x pulls data from SCCM, the service will use a .CONFIG file to define all the queries it will use to pull data. In general that file is either SCCM.LIGHT.EXTRACTOR.CONFIG or SCCM.FULL.EXTRACTOR.CONFIG. The LIGHT file has the minimum queries needed to normalize SCCM and make it useful. The FULL file includes a bunch of extra SCCM data for those people who want to just pass extra SCCM data on through and report off Normalize and SCCM data together instead of having to pull some data from Normalize and some data from SCCM.
So, that’s what pulls data from SCCM, but what it actually pulls is data like Add/Remove Programs data (v_GS_Add_Remove_Programs_Data and v_GS_Add_Remove_Programs_64_DATA) as well as software inventory data (v_GS_SoftwareFile and v_ProductFileInfo) and software metering data (v_MonthlyUsageSummary). I’m not privy to the actual algorithm Normalize uses to match SCCM data to the catalog, but I know it tries to match the ARP, software inventory and metering data to products in the catalog. Well, I don’t know about you, but in order to keep our SCCM database reasonably sized, we’re only inventorying a few EXEs with Software Inventory, we’ve only got rules on a handful of EXEs for Software Metering and many of our packagers in the past have changed the Add/Remove Programs names of software they repackage so it no longer matches anything in anyone’s catalog. What this all means is there are actually fewer data points than there should be to match to the Normalize catalog.