Hi all,

I was asked to explain parts of the script below, and in case there
are others that are interested, I've decide to make my reply public.
Also, I wrote that 7.0+ was needed to run the script below, but that
is incorrect. I think you just need 6.0 exp 1 or later.

...
> PARM acct=!hpaccount
> # This script displays all jobs that match the specified account
> #
--> CI comments can begin with a '#' just like a unix shell.

> setvar _sj_acct ups("!acct")
--> I upshift the account name for display purposes only. The JOBCNT
    function works fine with lowercase.

> if setvar(_sj_cnt, JOBCNT("@j:@.!_sj_acct",_sj_jobs)) = 0 then
--> I invoke the relatively new JOBCNT function. Try HELP JOBCNT on
    your system to see if you have it.  There is an article on JOBCNT
    on our Jazz webserver at:
http://jazz.external.hp.com/papers/Communicator/6.0/exp1/ci_enhancements.htm
l
(watch line wrap)
    So, if there are more than zero jobs currently running (or waiting or
    scheduled) then the return from JOBCNT will be > 0 and we want to
    process them.  Last, the matching job numbers (J123, etc) are stored
    in the CI variable names "_sj_jobs" with one space in between each
    job number.

>    echo No jobs found in the !_sj_acct
>    return
> endif
> echo !_sj_cnt jobs in the !_sj_acct found...
> setvar _sj_i 0
> while setvar(_sj_job,word(_sj_jobs," ",setvar(_sj_i,_sj_i+1))) <> "" do
--> parse the _sj_jobs variable extracting out each separate number (token)
    which in this case is delimited by a space. Before the extraction (done
    by the WORD function (help WORD)) increment the word index argument. So,
    we get the 1st job number, then the second (_sj_i+1) job number, then
the
    3rd... until we hit the end of the variable value (when the return from
    WORD is "" (empty)).

>    # display info about this job
>    setvar _sj_buf "!_sj_job "+JINFO(_sj_job,"JobUserAccountGroup")+" "+&
>                   JINFO(_sj_job,"FmtIntroDate") + " " + &
>                   JINFO(_sj_job,"JobState")
--> create a string variable that contains the matching job number,
    followed by the "Jobname,UserName.AcctName,GroupName", followed
    by the date the job was introduced, followed by the job's state:
    EXEC, SUSP, SCHED, INTRO. These fields will not be aligned in
    the output since each field is variable width. To get the fields
    in columns use the RPT function to insert a variable number of
    blanks based on the actual width of the data being echoed.

>    echo !_sj_buf
> endwhile
> if _sj_cnt > WORDCNT(_sj_jobs) then
--> The JOBCNT function may not be able to store all of the matching
    job numbers in the _sj_jobs variable, since a CI string variable
    is limited to 1024 bytes.  However, JOBCNT always returns the
    total number of matching jobs, not the number of entries that
    fit into the variable. To detect the (hopefully small number of)
    cases when the variable was not able to hold all of the
    matching jobs I use the WORDCNT function. This function returns
    the number of tokens in a string, so it will return the number of
    jobs that fit in the _sj_jobs variable.  If this differs from
    the function return from JOBCNT then you know that there has
    been truncation.  The solution to this is usually to further
    restrict the criteria JOBCNT uses to select matching entries.

>    echo some jobs could not be displayed, try again...
> endif


HTH,
 Jeff Vance, vCSY

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *