Shell Script

eluzi · 03-16-2006, 05:00 PM

Lads, i'd like to develop a shell script that would take a list of words in a file and told me what words are included there and how many times they're used... Gotta think...

Example:

world
window
linux
keyboard
linux
world
linux

Output:

world = 2
window = 1
keyboard = 1
linux = 3

Matir · 03-16-2006, 06:04 PM

Something like this should work:

Code:

#!/bin/bash

# Usage: script <FILENAME>

WORDLIST=`sort $1 | uniq`
for WORD in $WORDLIST ; do echo -n $WORD = ; grep -cE "^$WORD\$" $1 ; done

It's not terribly efficient, but it should work.

eluzi · 03-17-2006, 02:26 PM

Great ! Just to know the UNIQ command you helped me a lot, the only thing left now to do is to take something in reverse order, let me explain with an example to simplify:

www.visavale.com.br
www.vivario.org.br
www.vivifernandes.theblog.com.br

I want to take the top domain of these, so to the first is .COM and for the second it's .ORG, to the third is .COM. I'm trying to use the CUT command, using the dots as separator, but as you can see I'd have to take only the second field after the last dot...(coming from the end to the beggining) and then taking the first. So i'd have: .COM for the first and so on...
If I use CUT from the beggining and chose the 3 field i'd have .COM, .ORG, but in the third it would be 'theblog' and it's not what i want...
Is it possible?

Matir · 03-17-2006, 07:06 PM

Code:

sed 's/.*\(\.[a-z]{3}\)\.br$/\1/'

Should do it. Regular expressions are not my strength, so it may take some tweaking.