awk newbie

eremit · 03-25-2005, 11:26 PM

Hi,

my first post- woohoo!

I'm kinda new to awk scripting, but find it interesting and really powerfull when it comes to textstrings. But now I have encountered some problems; I'm building a database from a really old forum (html-files), and I try to separate posts from eachother. I know that every post starts with: From <username> <email>. Is there some way to get this process automated with awk? I would like every post to be saved in a separate file (eg. post1, post2, post3 etc.)

I have come this far:

awk -f <filename> /From/ { print $0 }

Also, I encounter some problems when the email field is present, it says illegal character in expression...

eremit

dustu76 · 03-26-2005, 02:39 AM

This is my first chance to welcome a first timer, so ... WELCOME

I'm no awk junkie, but what the heck... I tried by creating a sample master file (with incomplete records, empty posts thrown in for good measure)

Code:

[/home/soumen/tmp] $ cat post.txt
12349 234 askasd
12-3kasdweqj

From : a@b.com
asdfasfasf
 sadfasfas
 asdfasf

asdfasf

From : c@d.com
xsdfxsfxsf
 sxdfxsfxs
 xsdfxsf
xsdfxsf

From : e@f.com
From : g@h.com

The awk script is :

Code:

[/home/soumen/tmp] $ cat psplit.sh
#!/bin/bash

awk 'BEGIN {i=0; system("rm post?.txt") } 
        { if (/^From/) {
                ++i
        } else {
                fname="post"i".txt"
                print $0 >> fname
                close(fname)
        }}' post.txt
[/home/soumen/tmp] $

Output:

Code:

[/home/soumen/tmp] $ ./psplit.sh
[/home/soumen/tmp] $ for i in post?.txt ; do echo "--------$i--------"; cat $i; done
--------post0.txt--------
12349 234 askasd
12-3kasdweqj

--------post1.txt--------
asdfasfasf
 sadfasfas
 asdfasf

asdfasf

--------post2.txt--------
xsdfxsfxsf
 sxdfxsfxs
 xsdfxsf
xsdfxsf

[/home/soumen/tmp] $

I'm sure the awk gurus would have a more elegant awk script

HTH

eremit · 03-27-2005, 07:08 AM

Great stuff!

Just needed some minor tweaks to fit into my script

Thanks for giving me a helping hand

eremit