linux - grep, howto combine regular and invert match when invert pattern is part of regular pattern

Question

I'm facing the task of extracting all records that either contains the string "no issue" or does not contain the string "issue" and wonder if this can be achieved with a single call to grep. I know that this can be done using two calls, e.g.

grep "no issue" FILE > NEWFILE
grep -v "issue" FILE >> NEWFILE

However that will mess up the original order of the records so not an option as the order is of importance.

question from:https://stackoverflow.com/questions/66061528/grep-howto-combine-regular-and-invert-match-when-invert-pattern-is-part-of-regu

Answer 1

Using gnu grep:

grep -Pv '(?<!no )issue' file

RegEx Details:

• (?<!no ): Negative lookbehind that asserts a failure if we have word no followed by a space before current position
• issue: Match word issue

Or using gnu awk:

awk '/<no issue>/ || !/<issue>/' file