topic Re: Extract Numbers from specific text string? in SAS Programming

Extract Numbers from specific text string?

ybz12003 — Thu, 06 Jan 2022 20:38:08 GMT

Hello, I would like to extract numbers after text 'UGESTAGE'. Please advice, thank you. BTW, I haven't visited the community for a while, why I couldn't use the insert SAS code function. I apologize if the codes are not shown properly below. data Have; infile datalines delimiter='#'; input Comments : $200. ; datalines; UGESTAGE=34 5/7 WKS # UPREM, UGESTAGE = 32WKS # UGESTAGE =33.4 WKS # GASTROJEJUNOSTOMY DEPENDENCE, BILATERAL GRADE 4 VU ; run; data Want; infile datalines delimiter='/'; input UGESTAGE ; datalines; 34 / 32 / 33 / ; run;

Re: Extract Numbers from specific text string?

ballardw — Thu, 06 Jan 2022 21:39:04 GMT

You should be able to post program code in either a text box or code box opened by clicking on the <> or "running man" icons that appear above the message window (7th and 8th from left to right).

It is possible this is what you meant/wanted to post. Is the first data step correct? or should there be 3 lines of USESTAGE=?

data Have;
 infile datalines delimiter='#';
 input Comments : $200. ;
 datalines;
 UGESTAGE=34 5/7 WKS # UPREM, UGESTAGE = 32WKS # UGESTAGE =33.4 WKS # GASTROJEJUNOSTOMY DEPENDENCE, BILATERAL GRADE 4 VU 
;
 run;
 data Want;
 infile datalines delimiter='/';
 input UGESTAGE ;
 datalines;
 34 / 32 / 33 / 
;
 run;

Are you sure that Want data set contains what you really want? It only has the value of 34 for one record.

Re: Extract Numbers from specific text string?

HB — Thu, 06 Jan 2022 23:45:17 GMT

I'm really not good at all with RegEx but I would like to be. This may work with some of your data until you have a different pattern that breaks it. But maybe it will get you started.

data extracted;
	set have;
	pattern = prxparse ("/(?<=[=|= ])\d\d(\.[0-9]{1,2})?/");
	call prxsubstr (pattern, comments, start, length);
		if start gt 0 then do;
			ugestage = substr(comments, start, length);
			output;
		end;
	keep ugestage;
run;

proc print data =  extracted;
run;

gives me

Obs	ugestage
1	34
2	32
3	33.4

Edit: Ballard's question about 33.4 is very important.

I updated the RegEx to return 33.4

Re: Extract Numbers from specific text string?

ybz12003 — Fri, 07 Jan 2022 03:05:48 GMT

Re: Extract Numbers from specific text string?

ybz12003 — Fri, 07 Jan 2022 03:09:05 GMT

Sorry, I don't see any <> or 'running man' icon.

Re: Extract Numbers from specific text string?

Patrick — Fri, 07 Jan 2022 03:17:13 GMT

Re: Extract Numbers from specific text string?

ybz12003 — Fri, 07 Jan 2022 03:32:00 GMT

data Have;
      infile datalines delimiter='#';
  input Comments : $200.  ;
datalines;
	UGESTAGE=34 5/7 WKS #
	UPREM, UGESTAGE = 32WKS #
	UGESTAGE  =33.4 WKS #
	GASTROJEJUNOSTOMY DEPENDENCE, BILATERAL GRADE 4 VU
;
run;

data Want;
      infile datalines delimiter='/';
  input UGESTAGE  ;
datalines;
	34 /
	32 /
	33 /
	
;
run;

I use Chrome, all the function icons are shown again. I think the previous IE was too old.

Re: Extract Numbers from specific text string?

Tom — Fri, 07 Jan 2022 04:34:11 GMT

Do your example strings really start with four leading spaces like you have in your lines of in-line data?

If so your input statement that is using the normal $ informat will remove those leading spaces. To preserve the leading spaces you need to read the lines using $CHAR. Or copy the values from the _INFILE_ automatic variable after executing the INPUT statement.

Re: Extract Numbers from specific text string?

Ksharp — Fri, 07 Jan 2022 12:00:22 GMT

data Have;
 infile datalines delimiter='#';
 input  @'UGESTAGE' temp : $200. @@ ;
 UGESTAGE=scan(temp,1,'.','kd');
 datalines;
	UGESTAGE=34 5/7 WKS #
	UPREM, UGESTAGE = 32WKS #
	UGESTAGE  =33.4 WKS #
	GASTROJEJUNOSTOMY DEPENDENCE, BILATERAL GRADE 4 VU
;