Dracoro

Trying to figure out something that should be simple.

Have read a file and am looping through each line which is read into a $LINE variable OK.

Now, what I want to do is split that $LINE variable into it's constituent parts. The delimiter is a space.

Using ksh.

#There's gotta be a simple easy way to do this shirley.

dynamix

iTrader: (3)

http://uk.php.net/manual/en/function.split.php

will that do it?

TonyFlow

From memory, you would want to use awk or sed (havent done scripting for around 8 years though!

DrEvil

And then what, do you want to process each 'word'?

(sh/ksh)
for WORD in $LINE
do
echo $WORD
done

Or do you want to look at the 5 field?

echo $LINE | cut -d' ' -f5

Or you can use 'nawk' (or awk)

echo $LINE | nawk -F' ' '{print $5}'

Plus loads of other ways... hopefully the above might help..

chump

why not split each line into an array?

stevencotton

Because that'll use loads of memory if it's a big file. For this stuff you need perl rather than shell/awk/sed because it was designed for it:

#!/usr/bin/perl -w

use strict;

open FILE, '/path/to/file' or die $!;
while (defined (my $line = <FILE>)) {
my @fields = split /\s+/, $line;
# @fields is an array of the line split on one or more whitespace
... do whatever with it
}
close FILE or die $!;

jpor

iTrader: (1)

Here you go this should do what you want:

while read LINE
do
echo $LINE | awk '{print $1}' >> /tmp/fname
echo $LINE | awk '{print $2}' >> /tmp/sname
done < /tmp/test

As tested on AIX ver 5.3 ML1 in KSH

The script above is an example of having first name and surname in a ascii text file.

$1 = First Name
$2 = Surname

The script makes the assumption that you want to show 2 fields from each line in the ascii text file.

DrEvil

Agreed Perl is the best for this - is it me or does it actually not say what he wants to do with the 'fields' once he has them... Could make a difference to how complex or simple a script he needs...

if he is new to Unix, perl might be a little scary too... (Sorry don't know your background D.)

And finally - Steve stop showing off

stevencotton

s/.+e\/?.(e)n (.+t(...))/\/ \10\10Just another Perl hacker/;

Dracoro

Cheers chaps, got it working. Small file so the awk command did the job. I use Perl for large files but in this instance it would have been overkill.

jpor

iTrader: (1)

It also depends if the Pearl module is installed as well. Or running Pearl scripts would not be an option.

stevencotton

It's spelled perl You don't need to use any non-core modules for simple things, I only used pragmata.

jpor

iTrader: (1)

D'oh!

Although it depends on which platform he is coding on. I suppose for LINUX distros, Perl is included. But on some UNIX O/ses you have to make sure it has been installed with the BOS.

DrEvil

Yeah, AIX is different to Solaris / Linux in that way.

jpor

iTrader: (1)

Indeed it is