[oclug] sed, awk, perl, or ...?

[Raymond Wood]

On Wed, Feb 27, 2002 at 01:18:15AM -0500, Joe Burpee remarked:

> On Tue, 26 Feb 2002 at 22:57:53 -0600, Raymond Wood wrote:
> > o Open and search through a bunch of HTML files and locate all
> >   strings of the form 'image*_' (where the * represents either 2
> >   or 3 numeric characters).
> > o increment the 2 or 3 digit number by a value of '1'.
> > o optionally, do this recursively through several
> >   sub-directories.

> find /some/dir -name \*.html -print0 |
>   xargs -0 perl -pi.bak -e 's/image(\d{2,3})_/image@{[$1+1]}_/g'
> 
> This assumes you want to do all .html files below directory /some/dir.
> The -print0 and -0 avoid possible problems with weird characters like
> spaces in filenames.  Perl's -i option causes the files to be processed
> one by one, rather than concatenated; also the original files are backed
> up with suffix .html.bak.  The xargs utility is a handy way to break up
> argument lists into manageable chunks, and is probably safer and more
> efficient than alternatives such as a `for' loop.
> 
> For details see the man pages perlrun, perlre, and perlop; also info
> pages for find and xargs.

Joe, this 1-liner script works like a charm (but you doubtless knew that
already :)  I am in a state of amazement... 

Thanks to you and others for feedback received.

Cheers,
Raymond
-- 
Raymond Wood        XIST Information Services & Technology Inc.
Ottawa, Ontario     phone: (613)234-9621 X233  fax: (613)234-9564
http://xist.com     raymond.wood at xist.com    1-888-ASK-XIST