Recoding multi dimensional array variables in SAS

Question

I'm trying to recode a SAS array variable based on other variables condition. I am getting array subscript out of range error

      datasets Chk1 and Chk2 have the following data

      Chk1
       id Var1 Var2
       1   Y    N
       2   N    Y
       3   Y    Y
      Chk2
      id Var3 Var4
       1   N    Y
       2   Y    N
       3   Y    Y

      My desired output is
      Chk3
      id  Var1 Var2 Var3 Var4 a1 a2 a3 b1 b2 b3
       1    Y   N    N    Y    1           1 
       2    N   Y    Y    N       1     1
       3    Y   Y    Y    Y          1        1


      Here is the data merge with recoding variables

I need your help in tweaking do loop with incremental value.

      data Chk3;
       merge Chk1(in=a) Chk2(in=b);
       by id;
       if a and b;
       array wv(*) var1 var2;
       array wv2(*) var3 var4;
       array wv3(*) a1 a2 a3 b1 b2 b3;
       do i=1 to dim(wv3) by 3;
        if wv(i)='Y' and wv2(i)='N' then wv3(i)=1;
        if wv(i)='N' and wv2(i)='Y' then wv3(i+1)=1;
        if wv(i)=wv2(i) then wv3(i+2)=1;
       end;
       run;

All the arrays a single dimension.

Richard
– Richard

2019-05-14 01:57:34 +00:00
Commented May 14, 2019 at 1:57 — Richard
– Richard, Commented May 14, 2019 at 1:57

Richard · Accepted Answer · 2019-05-14 01:56:34Z

1

You are flagging an assertion of a bitwise combinatoric. The BOR and BLSHIFT bitwise functions will aid you greatly.

data have1; input
id Var1 $ Var2 $; datalines;
1   Y    N
2   N    Y
3   Y    Y
data have2; input
id Var3 $ Var4 $; datalines;
 1   N    Y
 2   Y    N
 3   Y    Y
run;

data want;
  merge have1 have2;
  by id;

  array have1bitCombo a1-a3;
  array have2bitCombo b1-b3;

  have1bitCombo[bor(var1='Y', blshift(var2='Y',1))] = 1;
  have2bitCombo[bor(var3='Y', blshift(var4='Y',1))] = 1;
run;

answered May 14, 2019 at 1:56

Richard

27.7k4 gold badges28 silver badges44 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

liuminzhao · Accepted Answer · 2019-05-14 01:30:25Z

1

below should work:

data Chk3;
    merge Chk1(in=a) Chk2(in=b);
    by id;

    if a and b;
    array wv(*) var1 var2;
    array wv2(*) var3 var4;
    array wv3(*) a1 a2 a3 b1 b2 b3;

    do i=1 to dim(wv2);

        if wv(i)='Y' and wv2(i)='N' then
            wv3((i-1)*3+1)=1;

        if wv(i)='N' and wv2(i)='Y' then
            wv3((i-1)*3+2)=1;

        if wv(i)=wv2(i) then
            wv3((i-1)*3+3)=1;
    end;
run;



Obs id var1 var2 var3 var4 a1 a2 a3 b1 b2 b3 i
1 1 Y N N Y 1 . . . 1 . 3
2 2 N Y Y N . 1 . 1 . . 3
3 3 Y Y Y Y . . 1 . . 1 3

answered May 14, 2019 at 1:30

liuminzhao

2,45519 silver badges28 bronze badges

Collectives™ on Stack Overflow

Recoding multi dimensional array variables in SAS

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related