On Wed, 31 Aug 2016 08:38:58 -0400, John Eells wrote: > >> BTW, is there a form of LISTCAT that lists both ALIAS and RELATED names? >Yes. > Feels like RTFM. Ah! I stumbled on ALL; GIYF.
In my (weak) defense, ALL does not appear in alphabetical order in that manual; it's well concealed in a list of mutually exclusive options. On Wed, 31 Aug 2016 07:36:28 -0500, Tom Marchant wrote: > >I was wrong about not needing to define the alias. If the symbol is not >defined, >there is no substitution, and the data set name remains unchanged as >SYS1.FOOBAR&NULL. > Bummer. It should produce an error message. But you'll get one (almost) soon enough. It would be a favor to customers if there were a standard system symbol evaluating to nothing with a name as short as possible. Can one use a symbolic alias with DISP=NEW? That would answer a FAQ, "How can I create a DSN incorporating a timestamp?" >And now I wonder. What happens if you specify SYMBOLICRELATE(SYS1.FOOBAR) >on the DEFILE ALIAS? The book doesn't say that you have to specify symbols to >be >substituted, only that you can. As it happens, I had to get someone else to >clean >up after my last test, so I'm reluctant to try. > I agree with your interpretation of the book. But long ago in these pages Peter Relson (IIRC) told me rather sternly that you must. (Everything not forbidden is compulsory?) RCF submitted. Thanks, gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
