Date:         Mon, 7 Oct 2002 15:19:51 +0100
Reply-To:     John Whittington <John.W@MEDISCIENCE.CO.UK>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         John Whittington <John.W@MEDISCIENCE.CO.UK>
Subject:      Re: How to count?
Comments: To: "Parent, David" <david.parent@CAPITALONE.COM>
In-Reply-To:  <E17yY75-00015K-00@coumxnn02.netbenefit.co.uk>
Content-Type: text/plain; charset="us-ascii"; format=flowed

At 09:39 07/10/02 -0400, Parent, David wrote (in part):

>I will concede that adding "retain count" may add some value to the
>programmer's eye when reading the program and gaining understanding.
>However, as initially written (and tested), I believe my implementation of
>the counter below will produce the same results with or without the retain
>statement.  Have you found differently?
>
>  -----Original Message-----
>From:   Cui, Jinrui [mailto:Jinrui.Cui@cshs.org]
>
>I just say to add "retain count", like
>
> > Data temp2;
> >    Set temp;
>         retain count;
> >    By x;
> >    /*Reset count to zero for each "x group"*/
> >    If first.x then count=0;
> >    Count+1;
> > Run;

Yes, you're certainly right.  The implied RETAIN which comes from coding
the summation statement "Count+1;" means that the same ';result' (value of
Count) will result whether or not one has an explicit RETAIN statement.  I
must confess that I often will include this 'unnecessary' RETAIN statement
- for the reason you mention.

If an explicit RETAIN had been included prior to the SET statement (which
is the coding practice of many folk), then there would be one slight
difference (if you count it as a difference!), in that the variable Count
would then be the 'first' variable in the output dataset, not the last one.

Kind Regards,


John

----------------------------------------------------------------
Dr John Whittington,       Voice:    +44 (0) 1296 730225
Mediscience Services       Fax:      +44 (0) 1296 738893
Twyford Manor, Twyford,    E-mail:   John.W@mediscience.co.uk
Buckingham  MK18 4EL, UK             mediscience@compuserve.com
----------------------------------------------------------------