I am currently studying for the LPI exams and have come up to regular expressions.
I have found studying these to be interesting and highly useful. However, I find 2 things a little strange:
1) lack of an AND boolean metacharacter. I would have expected to be able to link parts of a regexp using something like & or + (although I reliaze that these have different meanings) to perform AND operations. So for example (assuming # will become our AND metacharacter):
Look for text with cat AND dog on the same line:
At first I thought I could work around this with the following:
but this will only search out lines on which the characters cat appear before the dog characters; i.e. the search is word order sensitive, which I don't want.
So then I thought of a workaround using 2 greps, so:
grep 'cat' file | grep 'dog' file
Which works but feels to me like a bit of a hack. I just don't understand why there is no AND metachar.
2. More of a mystery to me, however, is the apparent lack of a NOT metachar.
I understand there is the ^ metachar which can me used is conjunction with square brackets to negate the appearance of certain character combinations in search strings:
but the frustrating thing about this is that the literal characters "d" "o" and "g" are individually considered and not "dog" as a whole.
So then I thought of adding -v (in the case of grep) to negate the entire regexp. This works great for this kind of search:
find all sentences that do not have dog OR cat in the sentence
grep -v '.*dog|cat.*
but if I want to do this kind of thing I am completely stuck:
find all sentences that do NOT have dog OR do have cat in the sentence
This is where i'd really like to have a ! metachar to negate things: eg:
but this doesn't seem to exist!
If anyone could help (especially with point 2) I would be very grateful.