#!/usr/bin/perl -w
use strict;
use DBI;
my $pattern = {pattern=>'A6 A9 A8 A5 A3'}; # widths of cols
my $dbh=DBI->connect('dbi:AnyData(RaiseError=1):');
$dbh->ad_catalog('T1','CSV','yourfile.csv');
$dbh->ad_catalog('T2','Fixed','yourfile.fix',$pattern);
$dbh->do("CREATE TABLE T2 AS SELECT * FROM T1");
__END__
[download]

Otherwise, you are, in effect, asking "please do this for me, for free, instead of your own work."

In fact, in this case, and being highly charitable about it, the "please" is only implicit!

• what's your longest field?;
• what perl operator could you use to pad those fields shorter than the longest to the length of the longest?
• and what would you still have left to do at that point?

Update: The last above is based on a guess that you want to create is classic "fixed length fields" -- where each field of each record is the same length, enabling use of pos to determine where one field ends and another begins. This, of course, may not be your intent... but if so refer to other's replies re ambiguity of your question, and to How do I post a question effectively?.

    s/.*/ /
[download]

There are many forms of "fixed length records". Binary, ascii,, mixed. But assuming you just want to space out the fields to fixed widths, you need to make two passes.

The first finds the maximum lengths of each field, by splitting using split or one of the csv modules as appropriate, and using length (and List::Util::max).

The second uses the maximum field widths from the first pass to format the data for output using either printf or pack.

You'll then need to decide whether each field should be left or right justified and what your overall record length should be and so on. And most of that will depend upon what you intend to do with the fixed width record file afterwards. Is it intended for use by another application? If so, what are the formatting requirements of that application?

There is enough there to get you started, but a clearer picture of what you are trying to do would help a lot.

I see many people suggesting a split and feeding the results through printf or sprintf to pad them to an appropriate length. While that works for the sample data above, I highly recommend using Text::CSV_XS to parse the comma-delimited file, because it will let you deal with more complex cases (and might actually be faster due to the XS code).

Also, while the printf family of functions works fine for this, isn't this particular application almost exactly what format and write were made for? Read perlform for more about using formatting to write out records.

What would be the best way to convert this comma delimited file into a fixed length one?

I'm not 100% sure, but is this what you are after?

#!/usr/bin/perl -w
 
use strict;
 
while (my $line=<DATA>) {
  chomp $line;
  my @f=split(",",$line);
  my $fmt = sprintf("%s","%10s" x 5) . "\n";
  printf $fmt,@f;
}
 
__END__
123456,12/01/2004,ABCDEFG,12345 ,XYZ
123456,12/01/2004,ABCDEFG,12345 ,XYZ
123456,12/01/2004,ABCDEFG,12345 ,XYZ
123456,12/01/2004,ABCDEFG,12345 ,XYZ
123456,12/01/2004,ABCDEFG,12345 ,XYZ
123456,12/01/2004,ABCDEFG,12345 ,XYZ
123456,12/01/2004,ABCDEFG,12345 ,XYZ
123456,12/01/2004,ABCDEFG,12345 ,XYZ
123456,12/01/2004,ABCDEFG,12345 ,XYZ
123456,12/01/2004,ABCDEFG,12345 ,XYZ
[download]

When that code is run you get:

    12345612/01/2004   ABCDEFG    12345        XYZ
    12345612/01/2004   ABCDEFG    12345        XYZ
    12345612/01/2004   ABCDEFG    12345        XYZ
    12345612/01/2004   ABCDEFG    12345        XYZ
    12345612/01/2004   ABCDEFG    12345        XYZ
    12345612/01/2004   ABCDEFG    12345        XYZ
    12345612/01/2004   ABCDEFG    12345        XYZ
    12345612/01/2004   ABCDEFG    12345        XYZ
    12345612/01/2004   ABCDEFG    12345        XYZ
    12345612/01/2004   ABCDEFG    12345        XYZ
[download]

Not sure what you mean by "fixed length" or what your criteria for the proper field width is, but the principle remains the same.

In future you'd get better responses if you are more specific about what your requirements are.

perl -pe 's/([^,]*),/ $1 . " "x(10-length($1))/eg' foo.txt
[download]