Date: Tue, 14 Apr 2009 17:59:01 -0400
Reply-To: Bob LaRue <bjlarue@SBCGLOBAL.NET>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Bob LaRue <bjlarue@SBCGLOBAL.NET>
Subject: Re: Problems with informats
As it turns out that is exactly what happened. One of the programs had a
local proc format with the same informat name but no CPROB and no 'other'.
Hence the crash. By the way, would you care to elaborate further on other
methods for batch submission. I am unfamiliar with SYSTASK.
Thanks a million,
On Tue, 14 Apr 2009 16:23:49 -0500, ./ ADD NAME=Data _null_;
>Could be that one of the INCLUDED programs redefines the INFORMAT.
>The programs that I am writing(trying to write) look very similar to
>the ones you are running with %RUNPROG. I "always" run my final
>versions as separate BATCH jobs. You can use a BAT or other type of
>SCRIPT, my favorite is a SAS program using SYSTASK.
>On 4/14/09, Bob LaRue <firstname.lastname@example.org> wrote:
>> OK, after extensive experimentation this afternoon I have become
>> that the method of invocation IS part of the problem. Our method of batch
>> submission looks like this:
>> %MACRO RUNPROG(PROG= ) ;
>> FILENAME a "&DIR\&PROG..SAS" ;
>> %INCLUDE a ;
>> %MEND ;
>> *%let DIR=V:\Statistical Programming\ST_208192
>> %let DIR=V:\Statistical Programming\ST_208192
>> options mprint mlogic symbolgen;
>> Once I do a submission like this, it ALWAYS messes up on the informat on
>> later runs in the same session, even if I rerun the program as a
>> submission. Restarting SAS in a new session solves the problem.
>> We have been submitting final code like this since I have been here in
>> company but this is the first study that I have seen these kinds of
>> despite using similar code before. Any ideas on what could be getting
>> corrupted? The proc format code is run in a separate file at the
>> of the session before the other programs are submitted.
>> Any thoughts on the above issues would be appreciated.