Date:         Mon, 12 Mar 2001 10:31:16 -0500
Reply-To:     Bob Burnham <robert.a.burnham@DARTMOUTH.EDU>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Bob Burnham <robert.a.burnham@DARTMOUTH.EDU>
Organization: Dartmouth College
Subject:      Re: Importing a . file dat

Hi Richard,

I'm not so sure that this is such a simple problem -- if it was just that
the data was spread over two lines, then that would be no trouble.  On the
other hand, if you always knew that the variable at the end of the first
line would wrap, then that would be easy to handle.  But I'm not sure there
is any 'simple' way when you don't know whether a variable has wrapped from
one line to the next consistently.  (Of course, if the SAS gurus know of
such a way then it would be fantastic if they could chime in.)

I can see a couple of different ways to handle it -- one would be to look at
the input line as a string and check to see if the last character is a
comma, which would mean that the field was complete.  Then you could parse
out the available variables and read the next line.  This sounds like a lot
of work to me (wow, am I lazy) when the alternative is just to simply
re-write the file so that each observation is one one line.  For example, if
I have a file called "names.txt" that looks like this:

Harry, Potter, 4 Privet
Drive
George, Bush, 1400
Pennsylvania Avenue

I can just read in each line of the file and then write it out to a
temporary location -- adding a CR only at the end of every other line.

%let LINES = 2;
data _null_;
  infile "names.txt" truncover;
  file "buffer.txt";
  length buffer $200;
  input buffer $200.;
  put buffer @;
  if(not mod(_N_,&LINES)) then put;
  run;
data addr;
  infile "buffer.txt" delimiter=',' dsd;
  length fname lname street $80;
  input fname $ lname $ street $;
  run;

Of course, you'll need to think about a couple of other things when you
actually do this -- the length of the each record in your file might mean
that you would have to change LRECLs and the width of the variable that you
are using for input.  All in all though, I think this might be the most
straightforward approach. . .

If you want to be really lazy (and impatient and full of hubris) then you
can reformat your file with Perl before reading it with SAS:

while(<STDIN>) {
  chomp;
  if ($. % 2) { print "$_ "; }
  else { print "$_\n"; }
}

Hope this helps (but there is probably a better way!!),

Bob

--
Bob Burnham
robert.a.burnham@dartmouth.edu
http://www.dartmouth.edu/~bburnham

"richard Simhon" <Richard.Simhon@CORNHILL.CO.UK> wrote in message
news:2710430F08DAD211A3F70008C71EEE6904333D01@s-cho-exch0001...
> All,
>
> This appears to be a simple problem but it's Monday!!
>
> I am importing a .dat file into SAS which uses a comma as a separator.
> Each record goes across 2 lines. A variable may not finish exactly at the
> end of the first line and may therefore be split onto the second line.
>
> Is there a way in the infile statement or elsewhere in the import
statement
> that I can get SAS to go to the second line and read the data for that
> record.
>
> Thanks in advance
>
> Richard Simhon
> Business Analyst
> Allianz Cornhill
> Tel +(44) 01483 552628