Date:         Wed, 3 Jun 2009 23:25:02 -0400
Reply-To:     Ya Huang <ya.huang@AMYLIN.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Ya Huang <ya.huang@AMYLIN.COM>
Subject:      Re: merge
Comments: To: fabianezte@GMAIL.COM

This can be done with a single sql step:

proc sql;
select *,monotonic() as origord, var(weig) as variance
from ani
group by treat,type
order by origord
;

 id  treat     age   weig  type   origord  variance
--------------------------------------------------------
       1  TA   333    365  P1201        1     924.5
       1  TA   453    603  P1801        2      12.5
     319  TA   236    408  P1201        3     924.5
     319  TA   345    608  P1801        4      12.5
       1  TS   300    355  P1201        5        32
       1  TS   441    593  P1801        6      2738
     318  TS   233    347  P1201        7        32
     318  TS   264    519  P1801        8      2738
       3  NI   299    393  P1201        9     480.5
       3  NI   362    589  P1801       10     840.5
     451  NI   220    362  P1201       11     480.5
     451  NI   276    548  P1801       12     840.5



On Wed, 3 Jun 2009 19:45:59 -0700, Bobs <fabianezte@GMAIL.COM> wrote:

>Hello,
>
>I have longitudinal data in long form (shown below), small example.
>I calculate the variance for each Weig by type into of each treat.
>Now, I want to make the merge to concatenate each value of variance of
>each Weig by
>id (subject) inside the treat, or I create a new column in the file so
>that id have the value of the variance for each corresponding Weig.
>
>data ani;
>input id treat $ age weig type $;
>datalines;
>1 TA 333 365  P1201
>1 TA 453 603  P1801
>319 TA  236 408 P1201
>319 TA  345 608 P1801
>1 TS 300 355  P1201
>1 TS 441 593  P1801
>318 TS 233 347 P1201
>318 TS 264 519 P1801
>3 NI 299 393  P1201
>3 NI 362 589  P1801
>451 NI 220  362 P1201
>451 NI 276  548 P1801
>;
>proc print;run;
>
>proc sort  data=ani out=weight_1;
>by treat type;
>run;
>proc print data=weight_1;run;
>
>proc means data=weight_1 var;
>class treat type;
>var weig;
>run;
>
>my database has to be as shown below
>
>data ani;
>input animal treat $ age weig type $ variance;
>datalines;
>1 TA 333 365  P1201   924.5000000
>1 TA 453 603  P1801     12.5000000
>319 TA  236 408 P1201   924.5000000
>319 TA  345 608 P1801  12.5000000
>1 TS 300 355  P1201    32.0000000
>1 TS 441 593  P1801    2738.00
>318 TS 233 347 P1201   32.0000000
>318 TS 264 519 P1801   2738.00
>3 NI 299 393  P1201    480.5000000
>3 NI 362 589  P1801    840.5000000
>451 NI 220  362 P1201   480.5000000
>451 NI 276  548 P1801   840.5000000
>; proc print;run;
>
>Anyone have a idea.
>Thank you in advance,