On Tue, 8 Feb 2022 18:27:04 +0000, Billy Ashton wrote:

>gil, I tried IEBGENER, but my test (copy to Unix using Gener, then copy 
>back to MVS file using Gener, then compare) failed with anything other 
>than text files (Ascii), but that might have been due to Path settings 
>or Filedata settings (too many iterations ago to remember).
> 
Due diligence.  Would DD FILEDATA=BINARY) make a difference?
Were you able to perform a similar verification using SORT?

>The delete is not necessary, and I could do the overwrite--it just 
>seemed cleaner to do it that way...I will review that again for 
>simplicity. And the /tmp directory suggestion about another subdir is a 
>good one-thanks!
> 
It's an MVS mindset:
    DISP=OLD fails if the data set does not exist.
    DISP=NEW fails if the data set exists.
    DISP=MOD appends rather than overwriting.
So the MVS tradition is to add an extra step with either
DISP=(MOD,DELETE) or DISP=(MOD,CATLG) to get a known state.
But with a UNIX mindset it's puzzling; PATHOPTS=(OCREAT,OTRUNC)
suffices.

>Finally !cp is not a negate...in sftp, the bang means to run a shell 
>command within the sftp active environment. So here, !cp is running a 
>shell Copy command to copy the MVS file to Unix (still trying to work 
>that out with return code testing).
>
Ah! I've known non-IBM FTP clients that use "|" for that.

-- 
Thanks,
gil

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Reply via email to