Yes, I’ve managed to code a very sophisticated structural matching algorithm that uses ontologies for input, a source and target DDL set of structures, and even (if you already have it) a partial cross-reference document.

RapidACE v2 is very smart in this manner.  I think IBM has a tool they sell for $150k or more that “builds cross-reference documents”.  But RapidACE v2 goes a few steps further.  The matching algorithm is unique, special – and will use a number of different components to determine a matching score!

Yes, it goes way beyond v1 – RapidACE v1 stuck to primary and foreign key declarations, RapidACE v2 actually uses sophisticated text parsing algorithms, along with naming conventions, and defined ontologies as guides.  Of course, for the really tough problems where no possible match can be found – you can also pre-define a sparse XREF (source to target) for that specific match; and voila! RapidACE v2 will take that in to account.

It’s a very cool process, the end result is a matching score – you can change the output threshold up or down (out of 100%) to determine what gets produced to the Excel cross-reference.  That’s right, RapidACE v2 actually READS and WRITES directly to Excel spreadsheets.

The other cool thing is: once you have “generated” your cross-reference, you can edit it in Excel – tighten it up, change it, modify it, then push it back in to RapidACE v2 in order to generate accurate ETL code for SSIS, SQLServer, or Informatica v9.x

See you on the flip side,
Dan Linstedt