using sed to parse large directory

mastro · 04-13-2006, 10:39 AM

Hello ppl,

I'm trying to use sed to parse a large directory of html files and remove a key word ex. 'wow' and then apply the change to the same html file, instead of writing it somewhere else.

So any help on how the syntax might look would be very much appreciated

Thanks

taylor_venable · 04-13-2006, 10:48 AM

How about this:

Code:

sed -e 's/wow/foo/g' -i .backup *.html

It will go through all .html files in the current directory, create a backup file with a .backup extension, then modify each file in place by substituting each instance of "wow" with "foo". Note you probably want to make the regex as precise as possible to avoid accidental unintentional changes (e.g. "s/man/boy" would turn "woman" into "woboy"). You can use

Code:

sed -f /path/to/file -i .backup *.html

to use the given file as your sed script.

(Disclaimer: these options are for BSD sed, but I believe they're the same for GNU sed.)

mastro · 04-13-2006, 10:58 AM

Thanks Taylor for the fast reply! Works like a charm