Date:         Wed, 21 Apr 1999 12:03:10 -0700
Reply-To:     "Berryhill, Timothy" <TWB2@PGE.COM>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         "Berryhill, Timothy" <TWB2@PGE.COM>
Subject:      Re: The WHERE statement does not work for rawdata
Content-Type: text/plain

Bob, I doubt most of us understand SAS and the C it is written in well
enough to be precise in this discussion.  I believe there is something
called a program data vector (PDV) which contains each observation as the
data step processes.  I believe the WHERE statement (or dataset option)
processes obs before they are moved to the PDV.  On the other hand, the
WHERE uses something from the PDV--if you drop a variable with a DROP= list
in the SET statement, you cannot test that variable in a WHERE.
 
A tiny quibble, the efficiency of the WHERE is the avoided I/O and
housekeeping.  I do not think WHERE is more efficient than IF.  If the data
happened to be such that no obs were deleted, I would not expect a WHERE to
select them all faster than an IF.
 
Tim Berryhill - Contract Programmer and General Wizard
TWB2@PGE.COM or http://www.aartwolf.com/twb.html
Frequently at Pacific Gas & Electric Co., San Francisco
The correlation coefficient between their views and
my postings is slightly less than 0
> ----------
> From:         Lambert, Bob[SMTP:Bob_Lambert@AFCC.COM]
> Reply To:     Lambert, Bob
> Sent:         Wednesday, April 21, 1999 11:01 AM
> To:   SAS-L@UGA.CC.UGA.EDU
> Subject:      Re: The WHERE statement does not work for rawdata
>
> I believe the WHERE statement processes conditions BEFORE the PDV is
> loaded
> and is therefore more efficient than the IF statement.  I would appreciate
> further clarification or debunkfication on this.
>
> Thanx,
>
> Bob Lambert
> > -----Original Message-----
> > From: Adolf Quast [SMTP:adolf-a.quast@DB.COM]
> > Sent: Wednesday, April 21, 1999 7:25 AM
> > Subject:      Re: The WHERE statement does not work for rawdata
> >
> > Hi,
> >
> > the WHERE statement is very efficient because it uses the information
> > stored in
> > the PDV (program data vector) of a SAS-dataset. The PDV is created when
> a
> > SAS-Dataset is built. The PDV contains information about field length
> and
> > type.
> > Your rawdata do not have a PDV, so they have to be read via INFILE and
> > INPUT.
> >
> > Greetings
> >      Adolf Quast
>