Counting total number of matches with grep instead of just how many lines match

Does grep offer a way to count the total number of matches it makes? The -c option only returns the number of lines that matched the regex, but in this case I have multiple matches per line.

3 Answers

try this:

grep -o -E "your expression" file |wc -l

well, -E is just an example, it could be -P, -F etc. point is -o

test:

kent$ echo "abc xxx yyy"|grep -cP "[a-z]{3}"
1
kent$ echo "abc xxx yyy"|grep -oP "[a-z]{3}"|wc -l
3

There is a -o flag which indicates that only the matched subsection of the line should get printed.

Use that in conjunction with wc -l:

grep -o "part of line" | wc -l

man grep explains it as well.

As an alternative to the other answers, using just grep:

grep -o "seach pattern" somefile.txt | grep -c ""

The -o in the first grep outputs each match, and just the match - not the entire line (unless the entire line IS the match, of course). The -c in the second grep then counts 'em.

It's a few more characters to type (like, 4 or 5), but I find it easier to remember.