SED Mass find and replace

I'm trying to do a mass find and replace using sed. I'm not very good with sed so I don't know how to make it find "" and replace it with "" in a few thousand files.

Here's what I have so far: sed -e "s/\

But I dont know where to go from here.

Who is Participating?
GnsConnect With a Mentor Commented:
You could wrap it up in a for-loop, or just use the -i option.... This is from the manpage:
       -i[suffix], --in-place[=suffix]
              edit files in place (makes backup if extension supplied)
So you could for example do
sed -i -e 's/' filenames
sed --in-place=".bup" -e 's/' filenames
Perhaps safest to test with one or two files to see that it really does what you expect, and then using the backup suffix method when you do 'em all... provided you have the space (of course:-).

-- Glenn
Just a note... many versions of sed do not have -i.   I believe it appears as a GNU extension in more recent versions only... If you find yourself on a system with a different sed implementation that doesn't have the -i option, and still want a one-line command, you can try perl, also  

I.E.  The perl way to do this would be to issue the command:  perl -i'.bak' -pe 's/cpdigest\.com/' *
(In that case you have a perl regular expression instead of a POSIX regular expression, also)

If you want to do the replacement recursively, you need a shell script.
Here's one way:


for i in `find /path/to/start/from -type f` ; do
       mv $i $i.tmp$$
       sed -e 's/cpdigest\.com/' < $i.tmp$$ > $i
       mv $i.tmp$$ $i.bak

You can also use find options like   -name \*.txt

for example, if you only want to alter files that contain that pattern in the filename
Hi wgordy,
go for Gns solution but i would use find instead of a for loop(for loops always has problem with strange file names etc.), like this:
find /path/to/files -type f -exec sed -i -e 's/' {} \;
but this is kind of dangerous so test with this first:
find /path/to/files -type f -exec sed -e 's/' {} > {}.kalleballe \;
this way you ca take a look at some files named *.kalleballe and see if it worked, and then use the command line above.

Well Mysidia, this is the Linux TA... It's one of the very few places where we can safely rely on the GNU-ish features being there;-).
And sure, you need a version 4 sed... But this isn't _that_ brand new exactly. Most modern distros will have moved to this version by now.
Further, if the feature showed to be lacking, I was planning to show some further examples along the same vein you and Rob did... (You just beat me to it (I'm much for dialog with the questioner;)...:-) Anyway... Thanks for doing them!

Oh, Rob, minor point... When we're talking thousands of files I tend to prefer using the xargs strategy WRT find -exec, simply to reduce the number of fork/execs that need be done ... so that transforms your first example to
find /path/to/files -type f -print | xargs sed -i -e 's/'
... Your second example doesn't lend it self to that though (for obvious reasons:-) but is fairly close to the
find /path/to/files -type f -print | xargs sed --in-place=.bup -e 's/'
which is inverse and fairly easy to "undo" (execise left to the reader:-):-)

-- Glenn
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.