topic Re: SUBSTRING check in SAS Programming

SUBSTRING check

Ivy — Wed, 13 Apr 2016 02:10:03 GMT

Hi,

I have two variable a & b, I would like to check whether b string is in a string .

index(a, b ) could not pick up some values, for examples:

data _null_ ;

a = "Bevacizumab,Cisplatin,Pemetrexed Disodium" ;

b = "Bevacizumab,Pemetrexed Disodium" ;

c = index( a, b ) ;

put c ;

run ;

Should I need to seperate the B string and then use index function or whether there is another function to check this ?

Thank you !

Re: SUBSTRING check

Reeza — Wed, 13 Apr 2016 02:20:47 GMT

You'll have to separate the strings.

The INDEX() function looks for an exact match, not for each of the words.

You can try PRX functions or put the words into a temporary array and use that.

It depends on how generalized your problem is, is the B changing per row, or fixed across the data set?

Re: SUBSTRING check

Ivy — Wed, 13 Apr 2016 02:23:41 GMT

Thank you, Reeza .

B is changing per row, some B string has one part, some has two or three parts. Maybe I need to separate them. Thank you again !

Re: SUBSTRING check

Reeza — Wed, 13 Apr 2016 02:33:17 GMT

Yes, you need to separate them. Use the scan function.

If you make the data long, you can change it to a SQL query which should be easier to code as well.

Re: SUBSTRING check

stat_sas — Wed, 13 Apr 2016 03:47:06 GMT

data want(drop=search_string i flag);
set have;
flag=0;
length search_string $30;

do i=1 to countw(b,',');
    do search_string = scan(b,i,','); 
       if index(trim(a), trim(search_string)) then flag+1;
    end;
end;

if flag=countw(b,',') then found='b in a';
else found='b not in a';

run;

Re: SUBSTRING check

PGStats — Wed, 13 Apr 2016 03:31:17 GMT

You couild always use regular expressions

data _null_; 
length pat $60;
a = "Bevacizumab,Cisplatin,Pemetrexed Disodium" ;
x = "Bevacizumab,Cisplatin,Pemetrexed Dissodium" ;
b = "Bevacizumab,Pemetrexed Disodium";
pat = prxchange("s/,/.*/", -1, b);
pat = cats("/", pat, "/i");
c = prxmatch(pat, a) > 0;
put c=;
d = prxmatch(pat, x) > 0;
put d=;
run ;

Will match if the substrings are in the same order.