=========================================================================
Date:         Fri, 19 Jul 2002 10:53:56 -0500
Reply-To:     goldbergj@MSNOTES.WUSTL.EDU
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         goldbergj@MSNOTES.WUSTL.EDU
Subject:      Re: Please help...SQL problem...
Content-type: text/plain; charset=us-ascii

It was opined:

I've spent a number of years
converting SAS data steps to SAS SQL and have found very few situations
where conversions in that direction did not produce more a understandable,
more robust, or more efficient program, and often all three at once.

But those few can be doozies.  I recently converted a SAS-SQL program back
to a data step because it used *very* complicated logical conditions to get
the effects of nested "if"s.  And yes, I'm familiar with the SQL "case"
construct.  The resulting structure was both hard to understand and
fragile.  Since I needed to modify it, I did the conversion for the sake of
sanity.  And it was well worth the effort.

In the process I also got rid of a buried bug.

I don't want to overdo this; SQL is often a straightforward way of doing
things.  But sometimes its lack of expressivity can make for devious
coding.  And its age shows sometimes, despite revisions.  Anyone else
remember Codd's old piece, "Fatal Flaws in SQL?"

Jonathan Goldberg
Missouri Alcoholism Research Center
Washington University School of Medicine
40 N. Kingshighway, Suite One
St. Louis, MO 63108