unix search/replace across files

[jdulberg]

I have a heap of files in a heap of directories that I need to remove a single line of text from. So far the only thing that I can find is the following but its not quite what I'm looking for.

perl -pi.bak -e s/oldstring/newstring/g filename(s)

I need to traverse the entire directory structure on my system looking at /home/vhosts/xyz.com/index.shtml for a string <!--#exec cgi=&quot;/cgi-bin/count.cgi&quot; --> and delete that entire line. All the files will be called index.shtml but xyz.com will be different in every instance.

Is such a command possible or do I have to do it manually?

Any suggestions are greatly appreciated!

Thanks
Jason

 

[J1mbo]

Something like:
find /home -name index.shtml | xargs perl -pi.bak -e 's/oldstring/newstring/g'

Taken from this FAQ:

http://www.tek-tips.com/faqs.cfm?spid=219&sfid=969&newpid=219

-jim

 

[jdulberg]

Thanks for your reply... I suppose next time I should look in the faq as well as doing a forum search. I tried the code from just below my test domain however it gave me a rather long error.

[vhosts]$ find domain.com -name &quot;index.shtml&quot; | xargs perl -pi.bak -e 's/<!--#exec cgi=&quot;/cgi-bin/count.cgi&quot; -->//g'

Bareword found where operator expected at -e line 1, near &quot;s/<!--#exec cgi=&quot;/cgi-bin/ax&quot;
syntax error at -e line 1, near &quot;s/<!--#exec cgi=&quot;/cgi-bin/ax&quot;
String found where operator expected at -e line 1, at end of line
        (Missing semicolon on previous line?)
Can't find string terminator '&quot;' anywhere before EOF at -e line 1.
xargs: perl: exited with status 255; aborting

Any ideas what went wrong??

Jason

 

[toolkit]

Your search string contains '/' characters which are confusing your 's/search/replace/g' call. In perl, you can use other characters instead of '/' in the call, for example:

[vhosts]$ find domain.com -name &quot;index.shtml&quot; | xargs perl -pi.bak -e 's@<!--#exec cgi=&quot;/cgi-bin/count.cgi&quot; -->@@g'

In fact for the example you have, where you wish to completely skip a line, try the following instead:

find . -name &quot;*.shtml&quot; | xargs perl -ni.bak -e 'print unless m[<!--#exec cgi=&quot;/cgi-bin/count.cgi&quot; -->]'

Note the -p argument has been changed to -n, which still loops through all the lines in your file, but does not print unless instructed in the -e argument.

Hope this helps

 

[jdulberg]

Toolkit, your second bit of code worked out perfectly... perhaps too perfectly. I think it would have been best to leave out the -ni.bak as there are now tons of index.shtml.bak files throughout the directory tree. How would I go about removing these files automatically?

Thanks again

Jason

 

[J1mbo]

find /directory -name index.shtml.bak | xargs rm

-jim

 

[jdulberg]

Thanks for your response J1mbo... Your command worked great!

I'm beginning to see how the format for these kinds of commands work.

Jason