On Mon, Feb 14, 2011 at 10:00 PM, Glenn Knickerbocker <[email protected]> wrote: > On 02/11/2011 06:27 PM, Rob van der Heij wrote: >> Somehow it sounds like you could do with your detail and master >> reversed and feed the references back in as masters instead > > OK, now that I've got my nested feedback loop working, I think I could > use some advice along these lines for collecting the nonmatches. > > At the bottom of the cascade, I now send the nonmatches back up to the > top. At some point I want to pick out the ones that *never* match.
So what makes them match later if they did not the first time? If it is because you want to lookup the retrieved master, then that could be done immediately and does not need to buffer everything. When you're building the master during the process (so you have forward references) it might be an option to feed all non-matched detail into the master of a 2nd lookup, and when the 1st lookup terminate feed its master table to the 2nd lookup as details.That also breaks the loop with undefined references. | Rob
