Okay, I've tried it all out now and I'd say that the parallel array solution is easy to do, easy to understand, and very traditional 4D. So, that's a comfortable option for any of us.
All the same, I think I'll use the solution from Justin Leavens and Rob Laveaux. This keeps everything in one tidy, well-organized container and (so Justin reports) takes advantage of 4D-specific implementation details to deliver better performance at scale. (Not sure how it compares with a binary array, but binary arrays presuppose sorted order, which isn't always something that you get for free in the first place.) In any case, this solution seems to take advantage of C_OBJECT more completely, so why not? In my case, I've got a bunch of objects and now they're pushed down another level because the objects are themselves values. If you look at the object as JSON, that's just a bit denser. No arguing that, but it seems worth it in a lot of cases. I tend to sweat these sorts of details for a bit, pick a solution and then apply it everywhere I have a similar problem. We've all got those habits, of course. So, I'm going for the key-as-key solution, where it applies, and not looking back...unless I do. Thanks to all for the help. ********************************************************************** 4D Internet Users Group (4D iNUG) FAQ: http://lists.4d.com/faqnug.html Archive: http://lists.4d.com/archives.html Options: http://lists.4d.com/mailman/options/4d_tech Unsub: mailto:[email protected] **********************************************************************
