He would use a DSN referback, not VOL=REF. And no, multi-volume isn't an issue. As C. Blaicher said above, the only difference is the system generates the DSN, and tries to get rid of the dataset automatically.
sas N.B. By "DSN" I mean "dataset name", not "dataset". Which I'm annoyed I have to say. On Tue, Oct 3, 2017 at 9:55 AM, Paul Gilmartin <[email protected]> wrote: > On 2017-10-03, at 07:39, Buckton, T. (Theo) wrote: >> >> Can somebody direct me to documentation on temporary data sets. I need to >> understand the characteristics and restrictions of these &&TEMP data sets: >> Space limits; >> Can it extend over more than 1 volume if so, would a multi-volume temp data >> set pose a problem if it is referenced again by a different DD statement in >> the same job step? >> > The JCL Ref. says: > > REF=*.ddname > ... > | VOL=REF obtains the volume serial numbers ... > > I assume use of plural implies multiple volumes are supported. > > | from the referenced ... earlier DD statement. ... > | No other information is retrieved by VOL=REF; in particular it does > not obtain > | the volume sequence number, volume count, or data set sequence number. > > Ugh. So the programmer must keep the data set sequence number on a > sticky note on a desktop. > > But I suppose that VOL=REF is superfluous for catalogued data sets and > that multiple temp data sets are rarely stacked on volumes. > > -- gil > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [email protected] with the message: INFO IBM-MAIN -- sas ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
