topic Re: Setting variable indicator based on matching character in another variable in SAS Procedures

Setting variable indicator based on matching character in another variable

munitech4u — Tue, 19 May 2015 14:36:03 GMT

I have a variable which is collection of strings.

I want to created another variable say x based on some condition like "if characters ('ABC,abc,aBc,)' found in variable y then set x="ABC_found" else "ABC_not found"

Re: Setting variable indicator based on matching character in another variable

RW9 — Tue, 19 May 2015 14:48:37 GMT

You haven't stated a question? Perhaps you want something like:

x=findw(upcase(YOUR_VARIABLE),"ABC");

X will then contain the first starting postion of ABC in the text.

Re: Setting variable indicator based on matching character in another variable

munitech4u — Tue, 19 May 2015 17:57:10 GMT

What if i wanted to check, like if 2 chars match ("ABC" or "CDE")

Re: Setting variable indicator based on matching character in another variable

RW9 — Wed, 20 May 2015 08:15:19 GMT

Well, that's a very different requirement than your previous request. There is the soundex function. Maybe able to do it with reg expressions also. The way I would do it though is to first decide on how many need to match. Then treat the string as an array and loop over it (I assume 2 chars need to match):

data want;

set have;

result=0;

do i=1 to length(strip(the_string));

if substr(the_string,i,1) in ("C","D","E") then result=result+1;

end;

run;

Then if result>=2 you know that at least 2 characters match irrespective of position.

Re: Setting variable indicator based on matching character in another variable

munitech4u — Wed, 20 May 2015 08:20:07 GMT

I cannot use substr function, as the character can occur at any position.

Re: Setting variable indicator based on matching character in another variable

RW9 — Wed, 20 May 2015 08:47:23 GMT

That is what the loop is for.

Iteration 1 = Is character at position 1 in the list of required characters, if so add 1 to total

Iteration 2 = is character at position 2 in the list of required characters, if so add 1 to total

Continue until end of string.

If total >= 2 - i.e if more than two required characters are found in the string any position, then there is a match.

Re: Setting variable indicator based on matching character in another variable

munitech4u — Wed, 20 May 2015 08:58:40 GMT

For some reason, result is always 0 coming for me. Is there anything we need to add for repeating over the number of rows?

Re: Setting variable indicator based on matching character in another variable

RW9 — Wed, 20 May 2015 10:24:34 GMT

Start a new topic. Post some test data, and what you want the output to look like.

Re: Setting variable indicator based on matching character in another variable

munitech4u — Wed, 20 May 2015 10:38:24 GMT

Nevermind here is what worked for me: I used the FIND with if condition and set the new variable to 1 if its found and repeated same in else for next keyword and so on.