Date: Tue, 9 Jan 2007 10:55:23 -0500
Reply-To: Chang Chung <chang_y_chung@HOTMAIL.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Chang Chung <chang_y_chung@HOTMAIL.COM>
Subject: Re: Using SCANOVER
On Tue, 9 Jan 2007 05:33:25 -0800, ajs2004@BIGFOOT.COM wrote:
>Suppose I have a really simple data step like this:
> infile datalines missover;
> input a b c;
> 1 2
>which gives me the data set I want (with one observation, and the value
>of C is missing).
>Now suppose I have this:
> infile datalines scanover;
> input @'A is' a @'B is' b @'C is' c;
> A is 1
> B is 2
>I get no observations, and the log shows
> NOTE: LOST CARD.
> a=1 b=2 c=. _ERROR_=1 _N_=1
>But what I want is to have that observation, even if the end of the
>file was reached without C being found - just as in the first example,
>I got the observation even though the end of the input line was reached
>without C being found.
>Is there an easy way to do this - I mean, without giving up the
>convenience of using SCANOVER and @'string'?
As far as I know, the answer is no. you can easily do something like below,
(1) one line, one variable
(2) the last line contains "****"
(3) one output obs with latest values or missing
retain a b c;
array v[*] a b c;
keep a b c;
do while (not index(_infile_,"****"));
do i = 1 to dim(v);
atStr = catt(upcase(vname(v[i])), " is");
loc = index(_infile_, trim(atStr));
if loc>0 then do;
value = scan(substr(_infile_, loc+length(atStr)),1);
v[i] = input(value, best.);
A is 1
B is 2
/* check */
proc print data=one;
/* on lst
Obs a b c
1 1 2 .