In days of yore, if the file was unblocked fixed length records, you could use BDAM to access by relative record number. In fact, you could even lie to BDAM if the file is RECFM FB and there were no short blocks in the middle. In that case, you gave BDAM the block number and deblocked the records yourself. I don't know if that is still possible as I haven't worked on an MVS system for over 25 years. Even if it is possible in z/OS, that is no guarantee that there is support for it in <MVS. Of that, I am completely ignorant.
Regards, Richard Schuh > -----Original Message----- > From: CMSTSO Pipelines Discussion List > [mailto:[email protected]] On Behalf Of Kris Buelens > Sent: Friday, April 02, 2010 10:14 AM > To: [email protected] > Subject: Re: read a data set backwards or nonsequentially? > > If one reads a plain MVS sequential dataset there is no way > than to read from the start of the file to the record you > want, this is not a PIPE restriction but inherent to the > organisation. CMS' EDF filesystem is in fact a kind of > indexed sequential dataset, that is, CMS can find record 567 > without reading all 566 preceding records. > So, I guess the piper didn't create a FILERAND for MVS as if > there is no saving on I/O's possible. > > IFIRC, things change when you use a VSAM ESDS dataset, it can > be used more like CMS' EDF filesystem. > > 2010/4/2 Glenn Knickerbocker <[email protected]> > > > I seem to be getting just far enough on MVS to start waxing > quixotic: > > Is there a way to read a data set into the pipeline out of > sequence, > > or is <MVS all we've got? If there is, is there a way to read > > backwards, or to get the record count to know where to start? > > > > ¬R > > > > > > -- > Kris Buelens, > IBM Belgium, VM customer support >
