为什么我从grep -oc“ foo”文件和grep -o“ foo”文件得到了不同的答案？ wc -l 哪个是对的？

Question

The question says it all really... I am trying to find the number of occurences of specific strings in a large file.
The answers are similar.... like 50 000 or so with the pipe and 49 000 odd with just the plain grep - oc...

Can anyone explain why I am getting different answers, and which one would be correct?

Thanks for the help.

Answer 1

grep -c counts only the matching lines whereas grep -o prints each matching part on a separate line which means that wc -l will print more "lines" than actually match.

As far as I can see, there is no effective difference between grep -oc and grep -c because the suppression of printing means that there is no resulting effect of -o any more.

For example, printing a three line "file" with one double match and one single match:

$ printf 'foo foo\nbar\nfoo\n' | grep -oc foo
2
$ printf 'foo foo\nbar\nfoo\n' | grep -o foo
foo
foo
foo

Answer 2

In --only-matching ( -o ) mode grep will sometimes output multiple lines for a single matching line.

alex@yuzu:~$ echo -e "foo\nbar\nbaz"
foo
bar
baz

alex@yuzu:~$ echo -e "foo\nbar\nbaz" | grep -o o  
o
o

But with --count ( -c ) it will count the number of matching lines.

alex@yuzu:~$ echo -e "foo\nbar\nbaz" | grep -oc o
1

So grep -oo | wc -l grep -oo | wc -l is counting all the matches, even if there is more than one match in a single line.