Date:         Fri, 11 Nov 2005 23:05:02 -0500
Reply-To:     "Richard A. DeVenezia" <radevenz@IX.NETCOM.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "Richard A. DeVenezia" <radevenz@IX.NETCOM.COM>
Subject:      Re: How to read a CSV file from right to left
Comments: To: sas-l@uga.edu

Rajat Mathur wrote:

> Actually, my raw data is something like this:
> ID,SOURCE,Var1,Var2......V275
>
> The Id is unique but unfortunately the SOURCE contains values which
> are comma separated. So my code to import to SAS go for a toss as for
> one observation there can be three source while in other there can be
> 20 source. And in my raw file the values for each variable is not
> enclosed within quotes("").
>
> However I don't want this variable (SOURCE) in my analysis.
>
> So I was wondering if I can read the data from V275 till V1 and drop
> the rest. By conventional method I can read ID. In these two obtained
> data I shall put a new ID which is nothing but the observation number
> and shall merge with this new ID to get my desired Dataset.
>
> Can anyone help me on this issue?

Try to decipher this fun sample.

data foo;
  retain id v1-v3 char1-char3;

  array cval[*] $32 char3-char1;
  array nval[*]   8 v1-v3;

  infile cards dsd dlm=',' _infile_=buffer1 ;
    file "NUL:" _file_=buffer2;

  input @1 id @;

  buffer1 = left(reverse(_infile_));
  input @1 (cval[*]) ($revers12.:) @;

  buffer2 = '';
  put @1 (char:)(',') @;

  buffer1 = buffer2;
  input @2 nval[*] ;

  drop char:;
cards;
1,a,b,c,3,4,5
2,a,z,5,4,3
3,a,b,c,d,e,f,12,34,56
run;

--
Richard A. DeVenezia
http://www.devenezia.com/