topic Re: empty string between look-behind and look-ahead in regular expression in SAS Programming

empty string between look-behind and look-ahead in regular expression

Quodly — Wed, 04 Dec 2019 12:27:11 GMT

data a;
  txt = "Name: Alfred"||'0D'x||"Age: "||'0D'x||"Height: 69";

  pat = prxparse("/(?<=Age: ).*(?=\r)/"); /* this does not work */
  call prxsubstr(pat, txt, beg, len);
  if beg then age = substr(txt, beg, len);
  output;

  pat = prxparse("/(?<=Age: )[^\r]*(?=\r)/"); /* this neither */
  call prxsubstr(pat, txt, beg, len);
  if beg then age = substr(txt, beg, len);
  output;
run;

instead of the age, the text until the next carriage return is selected. how can i assign the empty string without changing the look expressions?

Re: empty string between look-behind and look-ahead in regular expression

yabwon — Wed, 04 Dec 2019 12:55:42 GMT

Hi,

how about that:

data have;
  txt = "Name: Alfred"||'0D'x||"Age: 34"||'0D'x||"Height: 69"; output;
  txt = "Name: Anna"||'0D'x||"Age:  "||'0D'x||"Height: 69"; output;
run;


data a;
  set have;

  pat = prxparse("/(\bAge: ).*(\r)/");
  call prxsubstr(pat, txt, beg, len);
  if beg then age = scan(substr(txt, beg, len),2,":");
  output;
run;

All the best

Bart

Re: empty string between look-behind and look-ahead in regular expression

Jagadishkatam — Wed, 04 Dec 2019 13:04:54 GMT

Alternatively could try the prxchange

data have;
  txt = "Name: Alfred"||'0D'x||"Age: 34"||'0D'x||"Height: 69"; output;
  txt = "Name: Anna"||'0D'x||"Age:  "||'0D'x||"Height: 69"; output;
run;


data a;
  set have;
age=prxchange('s/(.*age:)(.*)(\sheight.*)/$2/oi',-1,txt);
run;

Re: empty string between look-behind and look-ahead in regular expression

Jagadishkatam — Wed, 04 Dec 2019 13:06:16 GMT

With your example

data a;
  txt = "Name: Alfred"||'0D'x||"Age: "||'0D'x||"Height: 69";

age=prxchange('s/(.*age:)(.*)(\sheight.*)/$2/oi',-1,txt);
run;

Re: empty string between look-behind and look-ahead in regular expression

s_lassen — Wed, 04 Dec 2019 13:58:58 GMT

The problem is not with PRX, it is with using a length of zero in a call to the SUBSTR function - the LEN variable is zero, but the SUBSTR function does not accept that - you get a note about "Invalid third argument to function SUBSTR" in both your calls.

Use the SUBSTRN function instead, that will correctly return an empty string.

Re: empty string between look-behind and look-ahead in regular expression

Quodly — Wed, 04 Dec 2019 14:21:39 GMT

great, thanks! with that i can still use those look expressions.