I am on Jbase now, and it refuses to build an index on any I-descriptor, so I can't test this. So, I'll take your word for it and file this under "You learn something new every day". Thanks, Charlie On 5/18/2009 10:27 AM, Scott Ballinger wrote:
On Sun, May 17, 2009 at 7:30 PM, Charlie Noah <[email protected]> wrote: Considering the nature of secondary indexes, and not just Universe, you should never count on an index built on a subroutine to be updated automatically. If Universe does this, it must be magic. ;-) The same goes for any dictionary definition that returns content from another file Charlie, I see the winkie, but still... It is perfectly appropriate to build indexes using I-descriptors and expect them to be updated automatically, e.g. to consistently format a name field for indexing, or to create compound keys, etc. Using I-descriptors (whether or not with SUBR calls) works perfectly and, in my experience, is no more error-prone than indexes on "native" fields. The fact that U2 secondary indexes are maintained automatically is the whole point: the system does that work for you. I agree that using SUBR or TRANS to utilize data from another file as part of the index is problematic. However, in that case the appropriate index scheme would be to maintain a separate inverted-key type file and maintain said file via a trigger. Frankly, it is a poor design that uses native U2 secondary indexes that need to be rebuilt each night. /Scott Ballinger Pareto Corporation Edmonds WA USA 206 713 6006 ------- u2-users mailing list [email protected] To unsubscribe please visit http://listserver.u2ug.org/ ------- u2-users mailing list [email protected] To unsubscribe please visit http://listserver.u2ug.org/
