Date:         Wed, 14 May 2003 10:40:30 -0700
Reply-To:     Wei <wei112@HOTMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Wei <wei112@HOTMAIL.COM>
Organization: http://groups.google.com/
Subject:      Re: suppressing quotes in export; breaking apart variables
Content-Type: text/plain; charset=ISO-8859-1

For your 2)

var1a = scan(var1, 1, ":");
var1b = scan(var1, 2, ":");

for your 3)

do while (index(line, "<") > 0);
        first = index(line, "<");
        len = Index(line, ">") - first + 1;
        word = substr(line, first, len);
        line = tranwrd(line, trim(word), ",");
end;



headmind@mindhead.com (simian) wrote in message news:<84843a13.0305132346.a2b5a9c@posting.google.com>...
> Hello all,
>
> I have a question about a kluge, and maybe some of you know a solution
> that will enable me to avoid the kluge entirely.
>
> 1) I'm trying to export a table as a colon-delimited file.  SAS
> defaults to putting double quotes around a value that contains a colon
> already.  How do I suppress this behavior?  Here's my current script:
>
> data _null_;
> set  EE.EE end=EFIEOD;
> file 'd:\ee_out.txt'
> delimiter=':' DSD DROPOVER lrecl=32767;
> format var1 $500. ;
> format var2 best12. ;
> format var3 $40. ;
> if _n_ = 1 then do;
> put     'var2' ':'
>       'var1a' ':'
>       'var1b'
>       ; end;
> do;
>       EFIOUT + 1;
>       put var2 @;
>       put var1 $;
>     ; end;
> run;
>
> 2) The point of doing the above is that I want to use the colons
> embedded in the values of var1 to break out var1 into two separate
> fields, var1a and var1b.  I had planned to do this by exporting the
> file as above, and then re-importing it as a colon-delimited file, but
> the double quotes foiled me.
>
> Is there an easier way to create new variables out of existing ones
> based on an internal delimiter in the data step without this
> import/export madness?  For example, creating out of a var1 value of
> 'alpaca:llama' new variables, var1a = 'alpaca' and var1b = 'llama'?
>
> 3) on a related note, how I got to 'alpaca:llama' in the first place
> was by having something like
> "<junk>alpaca<morejunk>llama<evenmorejunk>' in a long character
> variable.  I knew what the <junk> tended to look like so I killed it
> like so:
>
>    var1 = tranwrd(compbl(trim(var1)),"<junk>","");
>
> What I'd rather do is simply replace anything in < >'s with a
> delimiter (without knowing whether the contents of the < > are
> identically "junk" or not), and then delimit my file by those new
> delimiters.
>
> Any thoughts?
>
> Thank you!
>
> VJ
> headmind@mindhead.com