topic Re: Searching if a phrase exists between two variables in SAS Programming

Searching if a phrase exists between two variables

desas — Wed, 13 Sep 2017 23:40:59 GMT

Hi ,

I am working on a dataset where I am supposed to compare two variables to check if a phrase exists in one of the variables or not. I tried using index function and find function .Both the functions work if there a one-to-one match , I want SAS to tell even if there is that particular phrase in the variable .Ie , I want to check from my below example , variable A phrase is present in Variable b : Below is the example of what I want :

a	b	INDEX Function Result	Expected Result
Left ICA	Left ICA	1	Match
Left Middle cere - M1	Left Middle cere - M1	1	Match
Right Verte	Right Verte,Right Verte, mid cervi	0	Match
Right Middle cere - M1,Right Supra ICA	Right Middle cere - M1	0	Match
Midline Basilar trunk	Midline Basilar trunk,Right Verte	0	Match
Right Verte	Left Verte	0	No Matching Data
Left Middle cere - M1	Left Verte	0	No Matching Data

Is there an other SAS function or method can I use?

I hope I phrased my question correctly,

Thanks In Advance !!!

Re: Searching if a phrase exists between two variables

novinosrin — Wed, 13 Sep 2017 23:48:22 GMT

/*untested*/

Try eqt operator

proc sql;

create table want as

select *

from have

where a eqt b;

quit;

Re: Searching if a phrase exists between two variables

desas — Wed, 13 Sep 2017 23:56:38 GMT

Thanks for the reply ! , But it does not work. :(, Yet I am glad to learn that there is an 'EQT' operator in PROC SQL

Re: Searching if a phrase exists between two variables

Reeza — Thu, 14 Sep 2017 00:04:22 GMT

Can you show your code?

Did you use any of the modifiers to ignore case?

And I'm assuming you're using FINDW and INDEXW even though you say FIND/INDEX.

Re: Searching if a phrase exists between two variables

ChrisBrooks — Thu, 14 Sep 2017 00:12:24 GMT

i think this gives you the expected output - it searches both strings for the presence of the other

data have;
	length str1 $100 str2 $100;
	infile datalines dlm="~";
	input str1 $ str2 $;
	datalines;
Left ICA~Left ICA
Left Middle cere - M1~Left Middle cere - M1
Right Verte~Right Verte,Right Verte, mid cervi
Right Middle cere - M1,Right Supra ICA~Right Middle cere - M1
Midline Basilar trunk~Midline Basilar trunk,Right Verte
Right Verte~Left Verte
Left Middle cere - M1~Left Verte
;
run;

data want;
	set have;
	if find(strip(str2),strip(str1)) or find(strip(str1),strip(str2)) then flag=1;
	else flag=0;
run;

Re: Searching if a phrase exists between two variables

Astounding — Thu, 14 Sep 2017 01:22:35 GMT

Good idea. One more thing to consider: if upper vs. lower case should still be a match, and the 'i' modifier:

	if find(strip(str2),strip(str1), 'i') or find(strip(str1),strip(str2), 'i') then flag=1;

Re: Searching if a phrase exists between two variables

ShiroAmada — Thu, 14 Sep 2017 01:49:45 GMT

Try this:

data WANT;
  set HAVE;

MATCH=0;

NEW_A=compbl(tranwrd(tranwrd(A,",","-"),"-"," "));
NEW_B=compbl(tranwrd(tranwrd(B,",","-"),"-"," "));

if NEW_A=NEW_B then MATCH=1;

if MATCH=0 then do;
  COUNT_WORDS_NEW_A=countw(right(NEW_A),"");
  COUNT_WORDS_NEW_B=countw(right(NEW_B),"");

if COUNT_WORDS_NEW_A<=COUNT_WORDS_NEW_B then
  MATCH = NEW_A=substr(NEW_B,1,length(NEW_A);
else if COUNT_WORDS_NEW_B < COUNT_WORDS_NEW_A then
  MATCH = NEW_B=substr(NEW_A,1,length(NEW_B);
end;
run;

*Do google the SAS functions i used such as TRANWRD, COMPBL, COUNTW, SUBSTR, LENGTH.;

Hope it helps.

Re: Searching if a phrase exists between two variables

desas — Fri, 15 Sep 2017 17:46:33 GMT

Thanks All !!! This was very helpful !!!!!!