topic Re: Search for various strings with a variable in New SAS User

Search for various strings with a variable

Aidan — Fri, 07 Dec 2018 13:25:12 GMT

Hi there

What I want to do is search for a string within a large text string.

So I will have around 100 products for instance.

What I need to know is if any of these 100 products appear in a larger text string (which will have a lot of junk text but will have the product name in places) stored in a different table.

I am thinking I will need some sort of loop but not certain if their is an easier way to achieve this.

Any suggestions or helpful link or examples would be fantastic.

Thanks

Aidan

Re: Search for various strings with a variable

Ksharp — Fri, 07 Dec 2018 13:37:07 GMT

data want;

set have;

if prxmatch('/xx|yy|zz|aa|bb|cc/',long_string) then found=1;

else found=0;

run;

Re: Search for various strings with a variable

Ksharp — Fri, 07 Dec 2018 13:38:07 GMT

data want;

set have;

if prxmatch('/xx|yy|zz|aa|bb|cc/',long_string) then found=1;

else found=0;

run;

Re: Search for various strings with a variable

Ksharp — Fri, 07 Dec 2018 13:38:09 GMT

data want;

set have;

if prxmatch('/xx|yy|zz|aa|bb|cc/',long_string) then found=1;

else found=0;

run;

Re: Search for various strings with a variable

Aidan — Fri, 07 Dec 2018 13:43:14 GMT

Hi there

I dont think this will work for me, thanks all the same.

I found something below;

data _NULL_;
set &_INPUT1 end=last;
if _n_=1 then call execute('data want; set &_INPUT2; length wordsfound $50;');
call execute('if findw(LOG_NOTES,"'||strip(word)||'") >0 then wordsfound=catx(",",wordsfound,"'||strip(word)||'");');
if last then call execute('run;');
run;

This seems to work to some extent but it wont return any value for wordsfound, just set to null

Re: Search for various strings with a variable

Aidan — Fri, 07 Dec 2018 13:52:59 GMT

got it working using this thanks anyway 🙂

Re: Search for various strings with a variable

PaigeMiller — Fri, 07 Dec 2018 13:58:33 GMT

This part of the code from @Ksharp

'/xx|yy|zz|aa|bb|cc/'

can be created easily by PROC SQL. So assume the 100 product names are in a SAS data set named PRODUCT_NAMES (one product name on each row)

Then

proc sql noprint;
    select distinct name into :names separated by '|' from product_names;
quit;

%let names1=%str(%')/&names/%str(%');

data want;
    set have;
    if prxmatch(%unquote(&names1),long_string) then found=1;
    else found=0;
run;