Pattern matching/regular expression question 1

[teser]

How would I write something in Unix or Perl to change sentences or paragraphs in many web page documents. All the documents have some
similiar sentences and paragraphs. On occasion I need to change or
delete a specific sentence or paragraph. Any suggestions?

 

[pkiller]

you can use find and perl. find will search the files
you want. perl will change their contents.

for instance, suppose you want to change all
paragraphs &quot;<p>this is a paragraph&quot; to
&quot;<p>this is another paragraph&quot; on all your
html files under your current directory:

$ find . -name \*.html |
xargs perl -i -p -e 'BEGIN{undef $/} s/<p>this is a paragraph/<p>this is another paragraph/sg'

now it's a question of tunning the regular expression
inside s///. most of the times you will want to ignore
spaces, then you do:
s/<p>\s*this is ...\s*/<p>\n new ... \n/sg
also, you might want to change something inside your
tags (some attribute, for example):
s/(<img[^>]*src=&quotx.gif(&quot;.*?>)/$1y.gif$2/sg
(the $2 part wasn't really necessary, but makes it
more understandable)

cya

--
pkiller

 

[teser]

Thanks pkiller, that worked but I want to get an interactive script. I tried this:

#!/bin/ksh
echo enter data1
read data1
echo enter data2
read data2
find . -name p2 |
xargs perl -i -p -e 'BEGIN{undef $/} s/$data1/$data2/sg'


This doesnt change the data in file &quot;p2&quot;. Any suggestions?