Date:         Wed, 3 Dec 2008 10:19:04 -0600
Reply-To:     "./ ADD NAME=Data _null_," <iebupdte@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "./ ADD NAME=Data _null_," <iebupdte@GMAIL.COM>
Subject:      Re: Merge Many to one without repeat
Comments: To: Akramx <akram.chriai@gmail.com>
In-Reply-To:  <d1d69a06-a28a-4635-af65-446b7375a97b@j39g2000yqn.googlegroups.com>
Content-Type: text/plain; charset=ISO-8859-1

I think you can simply set all values to missing as in this example.
Any values from B would remain missing until a new obs from B is read.
 Values from A are updated on each observation.

data C;
   merge a(in=f1) b(in=f2);
   by ID;
   if f1;
   output;
   call missing(of _all_);
   run;

On 12/3/08, Akramx <akram.chriai@gmail.com> wrote:
> Hi everybody,
> I want to do a special merge.
> I have two tables A and B:
> data A;
>   input ID $ 1 Num 12.;
>   datalines;
> a 24
> a 65
> a 43
> ;
> run;
>
> data B;
>   input ID $ 1 color $5.;
>   datalines;
> a Red
> b Green
> c White
> ;
> run;
>
> the usual merge BY ID fot these tables is :
> code:
> proc sort data=a; by ID;run;
> proc sort data=b; by ID;run;
>
> data C;
> merge   a(in=f1)
>                b(in=f2);
> by ID;
> if f1;
> run;
>
> result (Table C):
>
> Obs    ID    Num    color
> 1     a      24     Red
> 2     a      65     Red
> 3     a      43     Red
>
> What I want is to merge only the first (or last) line By ID with the
> corresponding line in table B, like it's illustrated here:
> Obs    ID    Num    color
> 1     a      24     Red
> 2     a      65
> 3     a      43
>
> I can use : if not First.ID then Color=''; but you understand that
> it'a only a sample here, and in reality I have about 90 columns...
>
> thanks for your help.
>