> How do I cut a word out of a line with sed? The word end with a '_'. I > don't know, how to match word boundaries. I need "sample" > out of a sample_ line.
that blanks in the char class brackets are [ \t] (space and tab) you could expand the chars that can be part of a word, e.g. 0-9 cpuc% cat | sed 's/(^|[ ]+)[A-Za-z][A-Za-z_]*_([ ]+|$)/ /g' this is a test_ this test_ is another test_ this case too test_test shouldn't get deleted multiple test_ should also test_ work this is a this is another this case too test_test shouldn't get deleted multiple should also work
