How to extract a text part by regexp in linux shell? Lets say, I have file where in every line is an IP address, but in different position. What is the most simple way to extract those IP addresses using common unix command-line tools?
A:
I'd suggest perl. (\d+.\d+.\d+.\d+) should probably do the trick.
EDIT: Just to make it more like a complete program, you could do something like the following (not tested):
#!/usr/bin/perl -w
use strict;
while (<>) {
if (/(\d+\.\d+\.\d+\.\d+)/) {
print "$1\n";
}
}
This handles one IP per line. If you have more than one IPs per line, you need to use the /g option. man perlretut gives you a more detailed tutorial on regular expressions.
PolyThinker
2009-01-09 13:08:26
I do not know perl well... Could You write a full command?
softly.lt
2009-01-09 13:11:34
You could always write that in one line and use perl -n or perl -e to execute it but IMHO a small script like this is easier to handle, especially if you want to keep it and modify it later for other patterns.
PolyThinker
2009-01-09 13:24:16
+4
A:
You could use grep to pull them out.
grep -o '[0-9]\{1,3\}\.[0-9]\{1,3\}\.[0-9]\{1,3\}\.[0-9]\{1,3\}' file.txt
brien
2009-01-09 13:11:27
A:
You could use awk, as well. Something like ...
awk '{i=1; if (NF > 0) do {if ($i ~ /regexp/) print $i; i++;} while (i <= NF);}' file
-- may need cleaning. just a quick and dirty response to show basically how to do it with awk
A:
I usually start with grep, to get the regexp right.
# [multiple failed attempts here]
grep '[0-9]*\.[0-9]*\.[0-9]*\.[0-9]*' file # good?
grep -E '[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}' file # good enough
Then I'd try and convert it to sed to filter out the rest of the line. (After reading this thread, you and I aren't going to do that anymore: we're going to use grep -o instead)
sed -ne 's/.*\([0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\).*/\1/p # FAIL
That's when I usually get annoyed with sed for not using the same regexes as anyone else. So I move to perl.
$ perl -nle '/[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}/ and print $&'
Perl's good to know in any case. If you've got a teeny bit of CPAN installed, you can even make it more reliable at little cost:
$ perl -MRegexp::Common=net -nE '/$RE{net}{IPV4}/ and say $&' file(s)
JB
2009-01-09 13:35:48
+5
A:
Most of the examples here will match on 999.999.999.999 which is not technically a valid IP address.
The following will match on only valid IP addresses (including network and broadcast addresses).
grep -E -o '(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?) file.txt'
Omit the -o if you want to see the entire line that matched.
sjbotha
2009-01-09 13:46:53