topic sorting a dataset in SAS Procedures

sorting a dataset

michtka — Mon, 19 Mar 2012 15:04:57 GMT

Hi everyone, I got the next dataset with three variables, sorting by subjid and dpfdate:

txt

    1 SLOW INCRE
    2 ROUTINE DO
    3 ROUTINE DO
    4 ROUTINE DO
    5 ROUTINE DO
    6 ROUTINE DO
    7 ROUTINE DO
    8 ROUTINE DO
    9 ROUTINE DO
10 ROUTINE DO
11 ROUTINE DO
12 ROUTINE DO
13 ROUTINE DO
14 ROUTINE DO
15 ROUTINE DO
16 ROUTINE DO
17 ROUTINE DO
18 ROUTINE DO
19 ROUTINE DO

Obs DPFDATE SUBJID

    1 13DEC2006    10051001
    2 14DEC2006    10051001
    3 15DEC2006    10051001
    4 16DEC2006    10051001
    5 17DEC2006    10051001
    6 18DEC2006    10051001
    7 19DEC2006    10051001
    8 20DEC2006    10051001
    9 21DEC2006    10051001
10 22DEC2006    10051001
11 23DEC2006    10051001
12 24DEC2006    10051001
13 25DEC2006    10051001
14 26DEC2006    10051001
15 27DEC2006    10051001
16 28DEC2006    10051001
17 29DEC2006    10051001
18 30DEC2006    10051001
19 31DEC2006    10051001

I would like tocreate a dataset something like:

1 slow increase 13dec2006 10051001

2routine do 14dec2006 10051001

3routine do 31dec2006 10051001

I mean: sorting out by subjid and dpfdate but picking up only the first and the last record of every txt variable.

Thanks

sorting a dataset

Haikuo — Mon, 19 Mar 2012 15:17:54 GMT

I am not sure what you are really aiming for, but given the data as is, the following code can get what you want:

data have;

input txt &$11. DPFDATE :date9. SUBJID :$8.;

format dpfdate date9.;

cards;

SLOW INCRE 13DEC2006 10051001

ROUTINE DO 14DEC2006 10051001

ROUTINE DO 15DEC2006 10051001

ROUTINE DO 16DEC2006 10051001

ROUTINE DO 17DEC2006 10051001

ROUTINE DO 18DEC2006 10051001

ROUTINE DO 19DEC2006 10051001

ROUTINE DO 20DEC2006 10051001

ROUTINE DO 21DEC2006 10051001

ROUTINE DO 22DEC2006 10051001

ROUTINE DO 23DEC2006 10051001

ROUTINE DO 24DEC2006 10051001

ROUTINE DO 25DEC2006 10051001

ROUTINE DO 26DEC2006 10051001

ROUTINE DO 27DEC2006 10051001

ROUTINE DO 28DEC2006 10051001

ROUTINE DO 29DEC2006 10051001

ROUTINE DO 30DEC2006 10051001

ROUTINE DO 31DEC2006 10051001

;

data want;

set have;

by txt notsorted;

if (first.txt or last.txt);

run;

proc print;run;

Regards,

Haikuo

sorting a dataset

Linlin — Mon, 19 Mar 2012 15:45:46 GMT

/* I add a observation to your original data:

SLOW INCRE 23DEC2007 10051001 */

/* borrowed Haikuo's dataset */

data have;

input txt &$11. DPFDATE :date9. SUBJID :$8.;

format dpfdate date9.;

cards;

SLOW INCRE 13DEC2006 10051001

ROUTINE DO 14DEC2006 10051001

ROUTINE DO 15DEC2006 10051001

ROUTINE DO 16DEC2006 10051001

ROUTINE DO 17DEC2006 10051001

ROUTINE DO 18DEC2006 10051001

ROUTINE DO 19DEC2006 10051001

ROUTINE DO 20DEC2006 10051001

ROUTINE DO 21DEC2006 10051001

ROUTINE DO 22DEC2006 10051001

ROUTINE DO 23DEC2006 10051001

ROUTINE DO 24DEC2006 10051001

ROUTINE DO 25DEC2006 10051001

ROUTINE DO 26DEC2006 10051001

ROUTINE DO 27DEC2006 10051001

ROUTINE DO 28DEC2006 10051001

ROUTINE DO 29DEC2006 10051001

ROUTINE DO 30DEC2006 10051001

ROUTINE DO 31DEC2006 10051001

SLOW INCRE 23DEC2007 10051001

;

data temp;

set have;

n=_n_;

proc sort;

by txt n;

data want;

set temp;

by txt;

if (first.txt or last.txt);

run;

proc print;run;

Obs txt DPFDATE SUBJID n

1 ROUTINE DO 14DEC2006 10051001 2

2 ROUTINE DO 31DEC2006 10051001 19

3 SLOW INCRE 13DEC2006 10051001 1

4 SLOW INCRE 23DEC2007 10051001 20

Linlin