how to replace lots of text

acider · 11-10-2009, 05:57 AM

Hi everyone, I am new to Linux and unable to solve a problem. I have a html file and need to replace its text regularly using cronjob.

If example, the html file has

<font face=arial>Just a test.</font>
<font face=arial>This is a test.</font>
<font face=arial>This is just a test.</font>
<font face=arial>This is also a test.</font>
<font face=arial>This is only a test.</font>

I would require it to be replaced to the following regularly

<font face=arial>just a test.</font>
<font face=arial>this is a test.</font>
<font face=arial><b>this is just a test.</b></font>
<font face=arial><b>this is also a test.</b></font>
<font face=arial>this is only a test.</font>

What would be the easiest way to have a shell script replace everything from <font face=arial>J*only a test.</font> to a new chunk of text?

I was looking at sed examples but it seems to be good at changing only a few words. I am concerned about the file having special characters too, ie. <>?$& and not sure if sed could do the words + characters replacement well.

Pls help...

CmdoColin · 11-10-2009, 07:03 AM

sed is a stream editior, could possibly use that to modify something. You can use multiples in a line, for example:

A file called "stuff.txt" with:

I have two CD's for music
I have one CD for games
I have three CD's for a new distro

sed 's/CD/DVD/g' -e 's/distro/monkey/g' stuff.txt

Will replace the Word CD with DVD and distro with monkey, so afterwards the file "stuff.txt" becomes:

I have two DVD's for music
I have one DVD for games
I have three DVD's for a new monkey

You can also looks for only lines containing something:

sed -e '/three/s/new/naughty/g' stuff.txt

I have two CD's for music
I have one CD for games
I have three CD's for a naughty distro

Or you can put the lot together for:

sed -e '/three/s/new/naughty/g' -e 's/CD/DVD/g' -e 's/distro/monkey/g' stuff.txt

I have two DVD's for music
I have one DVD for games
I have three DVD's for a naughty monkey

Depending on what you want to achieve, could this help? Sed is pretty powerful, and there's a lot more too it depending on what you want to achieve. It is non-interactive, awk is possibly something else to use, it's a little more complex though. Definitely two things to look at.

themanwhowas · 11-10-2009, 04:30 PM

or you can use perls file handler abilities to take line x, do whatever to it then recreate the whole file using the modified line(s). Sed is probably the way to go but it's a bit complicated

sundialsvcs · 11-10-2009, 08:05 PM

There are several "power tools" which can be applied here.

sed is the "stream editor" which accepts a single stream of input, applies some filter or edit to it, and passes the result as its stream of output.
awk is a tool that is designed for processing more complicated files. Here you have a series of conditions that awk is to test for, along with rules that are to be applied to those records which meet each condition.
Perl is a full-fledged programming language that is especially suited to text manipulation. (And, well, just about anything and everything else that you could possibly think of...)

acider · 11-10-2009, 09:18 PM

Thanks for the replies.

I guess the sed examples won't work. It becomes too difficult to edit because there are a lot of text.

Ideally the perl or shell script should do the following:

1) check example.html file and look for this part

<html>*</h1>

the asterix means include everything that appears from <html> to </h1>, it does not matter what text is inside it, ultimately this whole part will be changed to something else, getting it to recognise the correct part (from <html> to </h1>) is more important

2) replace the above part into something else for example

<html>
<head>
<title>this is new title</title>
<script></script>
</head>
<body>
<h1>this is new text and is only a test</h1>

3) Text information in example.html is updated with the new text information from <html> to </h1>

Getting sed to look for lines won't work because the lines in example.html keep changing. As long as it could replace everything from <html> to </h1>, job is done.

What is the best way to tackle the above?

chrism01 · 11-11-2009, 12:38 AM

I'd definitely go with Perl for that.
http://perldoc.perl.org/
http://www.perlmonks.org/?node=Tutorials

acider · 11-11-2009, 12:42 AM

Wow thats too much for me to understand.

All I need is an example of a working script.

ghostdog74 · 11-11-2009, 12:51 AM

Code:

awk -F">" '{
  s=$1">" tolower(substr($2,1,1)) substr($2,2)">"
  print s
}' file

output

Code:

$ ./shell.sh
<font face=arial>just a test.</font>
<font face=arial>this is a test.</font>
<font face=arial>this is just a test.</font>
<font face=arial>this is also a test.</font>
<font face=arial>this is only a test.</font>

you did not say how the <B> comes about...so i will leave it as that.