You can use "Proc Format" to handle this. It is very easy.

在data步中该如何实现呢？

Here is the example for you. Remember, once you have your format built, you can use it in any data step and proc step. Let me know if you need further hep.


Proc Format;
    Value salerange
         low - 0=0
          1-3000 =1
          3001-6999=2
          6999-high =3;

Data Shoes (keep=Region Sales Band);
    set sashelp.shoes;
    Format band 1.;
    band = input(Put(Sales, salerange.), 1.);
Run;

下面首先生成一个数据集trial，包括我的关心的score和sev；

然后我用sql中的case expression生成一个新变量sev2,跟sev一样；

但是现在我想在data步中达到这个目的

DATA  TRIAL;
INPUT TRT $ CENTER PAT SEX $ AGE SCORE @@;
RESP = (SCORE GT 0);
/* RESP=0 (symptoms are absent), =1 (symptoms are present) */
IF (SCORE = 0) THEN SEV = 0; /* 0 No Symptoms*/
IF ( 1 LE SCORE LE 30) THEN SEV = 1; /* 1 mild Symptoms*/
IF (31 LE SCORE LE 69) THEN SEV = 2; /* 2 moderate Symptoms*/
IF (SCORE GE 70) THEN SEV = 3; /* 3 severe Symptoms*/
DATALINES;
A 1 101 M 55 5  A 1 104 F 27 0  A 1 106 M 31 35
A 1 107 F 44 21 A 1 109 M 47 15 A 1 111 F 69 70
A 1 112 F 31 10 A 1 114 F 50 0  A 1 116 M 32 20
B 1 105 M 45 20 B 1 108 F 44 65 B 1 110 M 32 25
B 1 113 M 61 75 B 1 115 M 45 83 B 1 117 F 21 0
B 1 120 F 19 55 B 1 122 F 38 0  B 1 125 M 37 72
;
run;

proc sql outobs=12;
   create table a1 as
   select score, sev,
          case
             when score eq 0 then 0
             when 1 le score le 30 then 1
             when 31 le score le 69 then 2
             else 3
          end as sev2
      from trial;
quit;
    

thank u,