On Tue, Feb 4, 2014 at 7:44 AM, Jousma, David <david.jou...@53.com> wrote:
> No, we have users sorting data with path names as input. Just ran a test. > Single path input, concatenated, all work fine, unless they add a DD DUMMY > between them. As a side note, DFSORT doesn’t do a good job with dynamic > sortwk allocations with PATH input because it cannot accurately determine > the input filesize. I had opened a ticket with IBM on this, and all I got > was "this is the way it is". My only option was to tell my developers to > hardcode sorkwk datasets in their job. > Normally, I praise the DFSORT people. But "it cannot accurately determine the input filesize" is _ignorant_. stat() on the file returns the size of the file in bytes. Of course, this doesn't for UNIX "files" which are not really files. Such as named pipes, or "/dev/fd/?" type "files" for open file descriptors. Which are not useful in batch anyway, so forget I mentioned them. > > //DFSORT EXEC PGM=SORT > //SORTIN DD PATHDISP=KEEP, > // PATHOPTS=ORDONLY, > // FILEDATA=TEXT, > // RECFM=VB,LRECL=255,BLKSIZE=25496, > // PATH='/etc/httpd1443.conf' > // DD DUMMY > // DD PATHDISP=KEEP, > // PATHOPTS=ORDONLY, > // FILEDATA=TEXT, > // RECFM=VB,LRECL=255,BLKSIZE=25496, > // PATH='/etc/httpd1443.conf' > > -- Wasn't there something about a PASCAL programmer knowing the value of everything and the Wirth of nothing? Maranatha! <>< John McKown ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN