Splitting Text lines?

0aniel · 11-21-2005, 06:04 PM

Hi there,

Just have a question regarding splitting text lines.
How do I get the following fixed strings;

The strings are fixed. I need it to
1.) Look for the first 8 digits and make split.(Global)

2.)Then look for the first 4 Letters and make split
What commands would I need to execute this in my Shell Script.

Thx a lot

PenguinPwrdBox · 11-21-2005, 06:44 PM

I would use perl.

Assuming that will always be the format, e.g.: 9600012301FXTF02FTXT03FFTF04TFXX05TXFF06TTTT
Also assuming these are being read from a file:

Code:

#!/usr/bin/perl -w

my $file = $ARGV[0];

open(FILE, "$file") || die "Could not open file: $!";

for(<FILE>){

    my $cwl = $_;
    $cwl =~ m/(\d{8})(\d{2}\w{4})(\d{2}\w{4})(\d{2}\w{4})(\d{2}\w{4})(\d{2}\w{4})(\d{2}\w{4})//g;

# Once this match is made, you can do something with $1, $2.....$6...which are all of your matches.
# Print them to a file, stdout, whatever....

}

chrism01 · 11-21-2005, 07:00 PM

If you know the number & length of the fields (ie fixed format), Perl is fine, but I'd use unpack() instead of a regex. It's easier on the eye.

PenguinPwrdBox · 11-21-2005, 07:11 PM

Was unfamiliar with unpack....
That rocks

shukla_chanchal · 11-21-2005, 09:02 PM

why cut will not work here ?
i think it is easier to do it with cut and script
although perl is best because its a strong programming langualge

chrism01 · 11-21-2005, 10:31 PM

Well, for a start you'd need a separate cmd line for each field with cut (I think).
Perl + unpack() will do it in 1 line.
Really though, use whichever you feel comfortable with.
The more complex your processing, the more I'd recommend using Perl.

0aniel · 11-22-2005, 04:38 AM

Hi there,

Its a bash script so
i need to use #unix commands.

PenguinPwrdBox · 11-22-2005, 11:00 AM

rewrite it in perl

0aniel · 11-22-2005, 02:54 PM

Cheers

evoGage · 11-30-2005, 03:08 AM

Quote:

Originally posted by PenguinPwrdBox
I would use perl.

Assuming that will always be the format, e.g.: 9600012301FXTF02FTXT03FFTF04TFXX05TXFF06TTTT
Also assuming these are being read from a file:

Code:

#!/usr/bin/perl -w my $file = $ARGV[0]; open(FILE, "$file") || die "Could not open file: $!"; for(<FILE>){ my $cwl = $_; $cwl =~ m/(\d{8})(\d{2}\w{4})(\d{2}\w{4})(\d{2}\w{4})(\d{2}\w{4})(\d{2}\w{4})(\d{2}\w{4})//g; # Once this match is made, you can do something with $1, $2.....$6...which are all of your matches. # Print them to a file, stdout, whatever.... }

How would you do that with sed commands instead for that example?