topic Re: Prxmatch to match a pattern of words in SAS Programming

Prxmatch to match a pattern of words

SK_11 — Mon, 21 Sep 2020 03:42:35 GMT

I am trying to find a pattern from a sentence e.g I am trying find " I am" and "live" and "with my parents" from "I am 20 years old and I live happily with my parents".

The code I tried is ;

data test;
sentence='I am 20 years old and I live happily with my parents';
pattern=prxparse('/(I am)\w*(live)\w*(with my parents)/i');
b=prxmatch(pattern,sentence);

run:

I am getting 0 value for b

Any help in this regards will be very much appreciated.

Re: Prxmatch to match a pattern of words

andreas_lds — Mon, 21 Sep 2020 05:31:52 GMT

Your aren't incorporating the blanks between the words you want to find and all other chars in the text. The following regex seems to solve the issue:

(I am)(\w*\W)*(live)(\w*\W)*(with my parents)

Re: Prxmatch to match a pattern of words

Shmuel — Mon, 21 Sep 2020 05:56:13 GMT

It seems that the string (\w*\W)* is used to skip any number of words.

Getting result b=1 means that the sentence contains all required substrings

in the defined order.

Next code is searching each sub string, if it exists in the sentence:

data _null_;
  sentence='I am 20 years old and I live happily with my parents';
  prxid = prxparse('/(I am)|(live)|(my parents)/');

  length unit $20;
  start=1;
  stop=length(sentence);

  do while (1);
     call prxnext(prxid,start,stop,sentence,pos,len);
     if pos=0 then leave;
    unit = compress(substr(sentence,pos,len));
    put unit=;
  end;
run;

Re: Prxmatch to match a pattern of words

SK_11 — Mon, 21 Sep 2020 06:03:03 GMT

Thanks a lot Andreas for an easy solution.