Ufdjrw

I'm searching through medical notes to capture all instances of a phrase, in particular 'carbapenemase producing'. At times this phrasing can occur > 1 time in a string. From some research I think PRXNEXT would make the most sense but I'm having difficulty getting it to do what I want to. As an example for this string:

if amikacin results are needed please notify microbiology lab at ext
for further testing the organism will be held until meropenem result
obtained by disc diffusion presumptive carbapenemase producing cre see
spmi for carba r pcr results not confirmed carbapenemase producing cre

From this comment above, I'd like to extract the phrases

presumptive carbapenemase producing

and

not confirmed carbapenemase producing

I realize I can't extract, I don't think, those exact phrases but some variation of it with a substring. The code i've been using I found here. Here's what I have thus far but it's only capturing the 1st phrase:

carba_cnt = count(as_comments,'carba','i');

if _n_ = 1 then do;
retain reg1 neg1;
 reg1 = prxparse("/ca[bepr]w+ prod/");
end;

start = 1;
stop = length(as_comments);
position = 0;
length = 0;

/* Use PRXNEXT to find the first instance of the pattern, */
/* then use DO WHILE to find all further instances. */
/* PRXNEXT changes the start parameter so that searching */
/* begins again after the last match. */

call prxnext(reg1, start, stop, as_comments, position, length);

lastpos = 0;
 do while (position > 0);
 if lastpos then do; 
 length found $200;
 found = substr(as_comments,lastpos,position-lastpos);
 put found=;
 output;
 end;
 lastpos = position;

 call prxnext(reg1, start, stop, as_comments, position, length);
 end;

 if lastpos then do;
 found = substr(as_comments,lastpos);
 put found=;
 output;
 end;

You are correct to use PRXNEXT for locating each occurrence of a regex match in a source. The regex pattern can be modified to use a group capture to search for an optional leading "not confirmed". The scenario for the least likely 'coder fail' is to focus loop and extract around a single call to PRXNEXT.

This example uses pattern /((not confirmeds*)?(ca[bepr]w+ prod)) and outputs one row per match.

data have;
 id + 1;
 length comment $2000;
 infile datalines eof=done;
 do until (_infile_ = '----');
 input;
 if _infile_ ne '----' then 
 comment = catx(' ',comment,_infile_);
 end;
 done:
 if not missing(comment);
 datalines4;
if amikacin results are needed please notify microbiology lab at ext 
for further testing the organism will be held until meropenem result 
obtained by disc diffusion presumptive carbapenemase producing cre 
see spmi for carba r pcr results not confirmed carbapenemase producing cre
----
if amikacin results are needed please notify microbiology lab at ext 
for further testing the organism will be held until meropenem result 
obtained by disc diffusion conjectured carbapenems producing cre 
see spmi for carba r pcr results not confirmed carbapenemase producing cre
----
;;;;
run;

data want;
 set have;
 prx = prxparse('/((not confirmeds*)?(ca[bepr]w+ prod))/');

 _start_inout = 1;

 do hitnum = 1 by 1 until (pos=0);
 call prxnext (prx, _start_inout, length(comment), comment, pos, len);
 if len then do;
 content = substr(comment,pos,len);
 output;
 end;
 end;

 keep id hitnum content;
run;

Bonus info: The prxparse does not need to be inside an if _n_=1 block. See PRXPARSE docs

If perl-regular-expression is a constant or if it uses the /o option, the Perl regular expression is compiled only once. Successive calls to PRXPARSE do not cause a recompile, but returns the regular-expression-id for the regular expression that was already compiled. This behavior simplifies the code because you do not need to use an initialization block (IF _N_ = 1) to initialize Perl regular expressions.