Date:         Thu, 17 Jul 2008 08:25:40 -0700
Reply-To:     karma <dorjetarap@GOOGLEMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         karma <dorjetarap@GOOGLEMAIL.COM>
Organization: http://groups.google.com
Subject:      Re: how to remove characters from observations when found at a
              specific place?
Comments: To: sas-l@uga.edu
Content-Type: text/plain; charset=ISO-8859-1

On 17 Jul, 13:34, Patrick <patrick.mat...@gmx.ch> wrote:
> That sounds like a case for "regular expressions" to me.
>
> The function PRXCHANGE should do the job:http://support.sas.com/onlinedoc/913/getDoc/en/lrdict.hlp/a002601591.htm
>
> Have a look athttp://support.sas.com/onlinedoc/913/getDoc/en/lrdict.hlp/a002288677.htm
> in case that you haven't used regular expressions before. It needs a
> bit of effort to understand how it works - but it's worth the time!
>
> Cheers
>
> Patrick

I found this a bit tricky to carry out with RegExp. You could use
something like the following, in this example leading 'S' ,'/' and 'L'
s are being removed:

data have;
input val $1-14;
cards;
/adgag/adgas/
LadgasgLfdaL
SdadfdSdsfasL
dagdagSadsfS
Ladsfasdfdf
;
run;

data want(drop=_:);
set have;
_len = length(val);
_start = val in:('/' 'L' 'S');
_end = reverse(strip(val)) in: ('/' 'L' 'S');
val = substr(val, _start+1, _len-(_start + _end));
run;

proc print;run;


1     adgag/adgas
2     adgasgLfda
3     dadfdSdsfas
4     dagdagSadsf
5     adsfasdfdf