Extract information between strings

[may0423]

Hi all, 

I have no idea how to extract information between strings.... for example I need the information Maple Grove, MN between <dt>Location</dt> and </dd>. (original info: <dt>Location</dt> <dd>Maple Grove, MN</dd>)

Thank you!

[Shmuel · Accepted solution]

You gave sample data as:

     <dt>Location</dt> <dd>Maple Grove, MN</dd>

assuming your data is in format:

   <dt>xxxxx</dt> <dd>yyyyyyyyy</dd>

then you can extract your info by:

data want;

  set have;

        pos1 = index(string, '<dd>') + 4;

        pos2 = index(string, '</dd>');

        info = substr(string,pos1, pos2-pos1 +1);

run;

[may0423]

thank you Shmuel for the immediate response! but the location format is not fixed. here are other examples:

<dt>Location</dt> <dd>Austin, TX</dd>

<dt>Location</dt> <dd>MD</dd>

[Kurt_Bremser]

---

@may0423 wrote:

thank you @Shmuel for the immediate response! but the location format is not fixed. here are other examples:

<dt>Location</dt> <dd>Austin, TX</dd>

<dt>Location</dt> <dd>MD</dd>

---

@Shmuels code will work in both examples, it does not use any fixed positions. Just try it.

[may0423]

Hi @Kurt_Bremser,

@Shmuel's code works! I might type in something wrong last night while adjusting the code........

Thank you both 🙂

[Patrick]

@may0423

Perl Regular Expressions offer a powerfull syntax when it comes to searching for patterns in a string.

It needs a bit of skill-up time before you can make use of it - but it's very worthwhile learning.

data have;
	input locat $50.;
	datalines;
<dt>location</dt><dd>Mapple grove,MN</dd> 
<dt>location</dt><dd>New york,NY</dd> 
<dt>location</dt><dd>New york,NY</dd> 
<dt>location</dt><dd>Austin, TX</dd> 
;
run;

data want(drop=_:);
  set have;
  length address $40;
  retain _prxid;
  if _n_=1 then _prxid=prxparse('/<dt>location<\/dt><dd>(.*)<\/dd>/oi');
  if prxmatch(_prxid,locat) then 
    address=prxposn(_prxid, 1, locat);
run;

http://support.sas.com/documentation/cdl/en/lefunctionsref/69762/HTML/default/viewer.htm#n1lru1b4uoo...

http://support.sas.com/documentation/cdl/en/lefunctionsref/63354/HTML/default/viewer.htm#p0s9ilagexm...

[may0423]

Hi @Patrick,

I am very new to SAS. Thank you for the notes! It's getting interesting 

[mnjtrana]

You could have used the SCAN function to extract the data based on delimiters, But SCAN function doesn't support 'word' delimiters. So you can try changing the delimitersor the <tags> and convert them to some symbols using tranwrd function and then can extract the information required.

Try running the below code:

data old;
	input locat $50.;
	datalines;
<dt>location</dt><dd>Mapple grove,MN</dd> 
<dt>location</dt><dd>New york,NY</dd> 
<dt>location</dt><dd>New york,NY</dd> 
<dt>location</dt><dd>Austin, TX</dd> 
;
run;

Data new;
	set old;
	length column1 $50. column2 $50.;
	column1=scanq(tranwrd(tranwrd(locat,'<dt>','^'),'</dt>','^'),1,'^');
	column2=scanq(tranwrd(tranwrd(locat,'<dd>','#'),'</dd>','#'),2,'#');
run;

output:

[rogerjdeangelis]

/* T100910 Extract or remove all text between two strings  */

inspired by
http://stackoverflow.com/questions/9721052/r-extract-part-of-string

and

https://goo.gl/KYV45R
https://communities.sas.com/t5/Base-SAS-Programming/Extract-information-between-strings/m-p/321238

also posted to SAS-L a couple of times

HAVE
====

ROGER DEANGELIS L15 NOVEMBER 21, 1963 49 SPACKENKILL ROAD

WANT
====

ROGER DEANGELIS 49 SPACKENKILL ROAD

OR

NOVEMBER 21, 1963


DETAILS
                    REMOVE ONLY
               ---------------------
ROGER DEANGELIS L15 NOVEMBER 21, 1963 49 SPACKENKILL ROAD
               ----------------------

                     KEEP ONLY
                    ------------------
ROGER DEANGELIS L15 NOVEMBER 21, 1963 49 SPACKENKILL ROAD
                    ------------------

SOLUTIONS


 ____
|  _ \ ___ _ __ ___   _____   _____
| |_) / _ \ '_ ` _ \ / _ \ \ / / _ \
|  _ <  __/ | | | | | (_) \ V /  __/
|_| \_\___|_| |_| |_|\___/ \_/ \___|


data _null_;
     retain compiled_pattern;
     length cleantxt $96;
     input txt & $100.;
     if _n_ =1 then do;
         compiled_pattern = prxparse('s/L15(.+?)49//');
     end;
     call prxchange(compiled_pattern,-1,txt,cleantxt);
     put cleantxt;
cards4;
ROGER DEANGELIS L15 NOVEMBER 21, 1963 49 SPACKENKILL ROAD
GREGG DEANGELIS L15 NOVEMBER 18, 1963 49 PENSACOLA L ROAD
;;;;
run;quit;
 _  __
| |/ /___  ___ _ __
| ' // _ \/ _ \ '_ \
| . \  __/  __/ |_) |
|_|\_\___|\___| .__/
              |_|

data _null_;
     retain compiled_pattern;
     length cleantext $96;
     input txt & $100.;
     if _n_ =1 then do;
         compiled_pattern = prxparse('s/L15(.+?)49//');
     end;
     if prxmatch(compiled_pattern, txt) then cleantext = prxposn(compiled_pattern, 1, txt);
     put cleantext;
cards4;
ROGER DEANGELIS L15 NOVEMBER 21, 1963 49 SPACKENKILL ROAD
GREGG DEANGELIS L15 NOVEMBER 18, 1963 49 PENSACOLA L ROAD
;;;;
run;quit;

[may0423]

Hi @mnjtrana,

Thank you very much for the code!