topic Re: Use prxmatch to match phrase from column values in SAS Programming

Use prxmatch to match phrase from column values

ABprint — Fri, 04 Nov 2022 00:57:57 GMT

I am trying to match a phrase using the prxmatch function by passing it a column value.

data long_string;
infile datalines delimiter=',';
informat txt $40. keyp1 $20. keyp2 $20. keyp3 $20. keyp4 $20.;
input txt keyp1 keyp2 keyp3 keyp4;
datalines;
That is a great way to end the day,great way,end,today,even worse
;
run;

How can I go through the 4 keyp variables and match the phrase?

Re: Use prxmatch to match phrase from column values

ABprint — Fri, 04 Nov 2022 02:14:51 GMT

This seems to work:

data first_try;
set long_string;
match1 = prxmatch(catt("m/(",lowcase(keyp1),")/oi"),txt);
match2 = prxmatch(catt("m/(",lowcase(keyp2),")/oi"),txt);
match3 = prxmatch(catt("m/(",lowcase(keyp3),")/oi"),txt);
match4 = prxmatch(catt("m/(",lowcase(keyp4),")/oi"),txt);
run;

I don't quite understand why my array is not working:

data final;
set long_string;
array keyp{4} $ keyp1-keyp4;
array match{4} match1-match4;
do i=1 to 4;
match[i] = prxmatch(catt("m/(",lowcase(keyp[i]),")/oi"),txt);
end;
run;

Re: Use prxmatch to match phrase from column values

PGStats — Fri, 04 Nov 2022 02:35:23 GMT

Well for one thing, the "o" pattern suffix instructs the matching function that the pattern is a constant, i.e. that it doesn't need to be recompiled at every invocation. Which is clearly not the case here.

BTW, the "i" suffix makes the matching case insensitive. So, no need for the lowcase() function call.

Re: Use prxmatch to match phrase from column values

ABprint — Fri, 04 Nov 2022 03:30:54 GMT

That fixed it, thank you for insight.