Question

I have an instance where certain URLs are getting malformed with equal signs and I need to be able to remove them.

An example broken URL:

http://www.go=ogle.com/search?ie=UTF

to be corrected to this:

http://www.google.com/search?ie=UTF

It can not simply replace the first occurrence of = because not all URLs are broken like this.

Is there a sed/awk or other regex way of deleting all instances of = only if they occur before a question mark?

Answer 1

use

sed -e 's~\(http://[^?]*\)=\([^?]*\)~\1\2~'

which basically says to strip one = character from within anything starting with http:// and then having anything but a ?.

edit looking at it again, this is a lot cleaner:

sed -e 's~\(http://[^/?]*\)=~\1~'