Date:         Fri, 4 May 2001 11:30:59 -0500
Reply-To:     Jonathan Goldberg <jonathan@MATLOCK.WUSTL.EDU>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Jonathan Goldberg <jonathan@MATLOCK.WUSTL.EDU>
Subject:      Re: Obsolete SAS statements and options
Comments: cc: John.W@MEDISCIENCE.CO.UK
Content-Type: TEXT/PLAIN; charset=US-ASCII

John Whittington <John.W@MEDISCIENCE.CO.UK> asks:

>At 17:41 01/05/01 -0600, Jack Hamilton wrote:
>
>>The data set history feature saved the source code itself.
>
>I have recently seen several references to this apparent feature of old
>versions of PROC CONTENTS, but I am a little intrigued.
>
>Exactly what 'source code' was saved, I wonder?  I would suspect that it
>will have been only the DATA or PROC step that actually created the
>dataset
>- i.e. the last stage in the process.  If that is the case then, at least
>with my programs, it could be very unhelpful in many cases, since that
>final step is often only the conclusion or bringing together of lots of
>code that went into creating the dataset in question.

I remember this.  The history stored was whatever created the data
set.  It was indeed "unhelpful in many cases" and I assume that is why it
was allowed to lapse.  However, if the data set was created directly or
indirectly by a data step it could be quite useful; I remember it saving
me tears on more than one occasion.

The "indirectly" case came about because several generations of source
code were stored.  So, if you had a data step, followed by a sort, you saw
both with the history option of proc contents.  But if it was the result
of proc sort with the out= option all you saw was the sort (IIRC).


Jonathan Goldberg
Missouri Alcoholism Research Center
Dept. of Psychiatry
Washington University School of Medicine
40 N. Kingshighway, Suite One
St. Louis, MO 63108
314-286-2212