non-greedy matching in sed

If you want to match a prefix in a string and stop at the n'th occurence of a delimiter in that string then you need non-greedy matching. This is how you would do it if the regex engine you are using doesn't support it natively, do it by negative character class.

Example: given a URI http://service.domain.tld/path1/path2/path3 and you only want service.domain.tld by splitting on /, then:

$ echo "http://service.domain.tld/path1/path2/path3" | sed 's@http://\([^/]*\).*@\1@'
service.domain.tld

Now what happens if you need the suffix after the delimiter? Just change the capturing parens:

$ echo "foo_bar_baz" | sed 's/[^_]*_\(.*\)/\1/'
bar_baz

Courtesy of: http://stackoverflow.com/questions/1103149/non-greedy-regex-matching-in-sed