I wanted to search a for all lines containing ERROR but not errors that contained the word "foo" (for example). The only way I could figure out to do it was:

grep ERROR myfile.log | grep -v foo

is there a way to do this with one grep command instead of two? One grep is faster than two, especially for large files like log files.

Use:

or:

Hi use
grep -r "ERROR" *

radoulov, isn't "fgrep ERROR file|fgrep -v foo "-the same as doing "grep ERROR file | grep -v foo", except it uses fgrep instead of grep? Is there some advantage for fgrep in this case?

The awk thing is pretty cool - I tried it and it works great. I'll have to look into awk more - I know nothing about it (yet!).

sanjustudy, I don't understand how grep -r "ERROR" * will do what I want. I thought "-r" was for searching directories recursively. Also, I don't see how it searches for "foo".

If you don't search for a RE, it should be faster: