> Charles: > > And some complain about the obtuse nature of "F" correlatives. :-) > > Bill
*Sigh* Yes, Bill, point taken. BUT! this is quicker to write and probably quicker in execution than most (all?) suggestions so far. I forgot to mention that because the target for TRANS() stays constant through the entire execution, the record is read only once & subsequent references are very cheap. When it comes to MV-handling, it's the ol' Power vs Simplicity trade-off. Because I-descriptors work on entire dynamic arrays, MV-handling can be more difficult looking, with all those mv-handling subroutines like EQS, IFS, REUSE(), etc.. Correlatives are designed to just specifiy how to act on a single value. For multiple values, its up to the query engine to repeatedly feed the next value to the correlative. I'll write a detailed explanation of the current example: TRANS(ctrlfile,ctrlid,-1,'X');EQS(@,REUSE(fld.in.question)) if anyone asks. cds > From: Stevenson, Charles > > Don't know about UD limits. Here's a workaround: > > Write a control record with the 251 different numbers that > you are searching for. > Then > SELECT file WITH EVAL "SUBR( '!EQS', TRANS( controlfile, > controlid, -1, 'X' ), REUSE( field.in.question ))" = "1" > > Something close to that, anyway. > It loops through the entire file once (the expensive part), > and does 251 comparisons for each record (not so sepenseive). > > cds ------- u2-users mailing list u2-users@listserver.u2ug.org To unsubscribe please visit http://listserver.u2ug.org/
