Hi Kastus,

You need ``find'' command with -exec. Find will get the files recursively, and exec will run sed to edit each file. I performed something similar against a CVS repository once, but IIRC I created a two-line wrapper script around sed (or I used vi) and called that script from -exec option of find.

Thanks Kastus - I'll look into doing something like that. -Tim -- ----------------------------------------------------------------- Timothy R. Butler Universal Networks Information Tech. Consultant Christian Web Services Since 1996 ICQ #12495932 AIM: Uninettm An Authorized IPSwitch Reseller tbutler@uninetsolutions.com http://www.uninetsolutions.com ============== "Information Powered by Innovation" ==============

Ok, that's just plain sexy. Off to the bookstore to get Learning Perl! Geordon ----- Original Message ----- From: "Tara L Andrews" To: Cc: Sent: Thursday, April 26, 2001 10:45 PM Subject: Re: [SLE] OT: I need to replace a certain string in a bunch of files

On Thu, Apr 26, 2001 at 09:35:09PM -0500, Timothy R.Butler wrote:

...

Hi everyone,

I'm looking for a way to replace every instance of a certain text string in about nine files. What's more, what I'm looking for preferably needs to replace these items recusively, as I have the aformentioned string in a directory, and it's subdirectory.

Any ideas? For some reason I have the feeling I am going to need to get to know "sed." <rant>If only I could get a version of UltraEdit for Linux, I wouldn't have this problem... :-( Unfortunately I can't find a text editor for Linux that has a multi-document interface/project interface, so it can do a "Replace in all Project files."</rant>

Alternatively, you could get to know perl. :)

find . -type f -exec perl -ni.bak e 's/old text string/new string of text/g; print' {} \;

(or \{\} \; at the end, to be perfectly safe)

Breakdown for those unfamiliar with the command(s): finds all files that are regular files (not directories or symlinks, etc.) in the current directory (.) and executes the perl command on each file found.

The perl says "On each line of a file, saving a backup in filename.bak, replace (switch, hence the 's') every occurrence of 'old text string' with 'new text string', more than once per line if you have to. Then print the line."

-tara

-- To unsubscribe send e-mail to suse-linux-e-unsubscribe@suse.com For additional commands send e-mail to suse-linux-e-help@suse.com Also check the FAQ at http://www.suse.com/support/faq and the archives at http://lists.suse.com

On Thursday 26 April 2001 10:51 pm, you wrote:

Ok, that's just plain sexy.

Off to the bookstore to get Learning Perl!

That would be a good choice! And once you get through that, I highly recommend Learning Perl's big brother - Programming Perl (the Camel book). -Tim -- ----------------------------------------------------------------- Timothy R. Butler Universal Networks Information Tech. Consultant Christian Web Services Since 1996 ICQ #12495932 AIM: Uninettm An Authorized IPSwitch Reseller tbutler@uninetsolutions.com http://www.uninetsolutions.com ============== "Information Powered by Innovation" ==============

Sorry, I could not resist:

...

find . -type f -exec perl -ni.bak e 's/old text string/new string of text/g; print' {} \;

$ find . -type f -exec perl -pe 's/old_string/new_string/g' {} \; is even shorter. The -p switch puts a loop-and-print around the command. The -i.bak switch is only needed when you want backup-files. As always in perl, there is more than one way to do it. Regards, Cees.

On Fri, 27 Apr 2001, Thorsten Haude wrote:

I like UltraEdit, but I like NEdit more. They are somewhat similar, NEdit is a quite Windows-user-friendly editor, but can be very powerful.

Get the latest version from ftp.nedit.org/pub/snapshot/ and see whether the Replace dialog is what you're looking for.

Or just install the nedit.rpm from series xap in Yast. Regards Ole