Question

I want a regular expression that will extract the words happy and good, both non greedy and both case insensitive.

@a = [" I am very HAppy!!", "sad today..", "happy. to hear about this..", "the day is good", "sad one", "sad story"]

It looks like this works with one word:

@z = @a.join.scan(/\bhappy\b/i)

But when I add in good it does not work as I expect.

@z = @a.join.scan(/\bhappy|good\b/i) 

Expect ( happy 2x and good 1x):

@z.size => 3

The result it gives me:

@z.size => 2

Answer 1

You should add parentheses around your alternation so that the \bs will apply to either happy or good as a unit:

\b(happy|good)\b

Then, you probably want to scan each element of the @a array rather than @a.join so a map and flatten are called for:

@a.map { |s| s.scan(/\b(happy|good)\b/i) }.flatten
# ["HAppy", "happy", "good"]

You could also use a non-capturing group:

\b(?:happy|good)\b

but it won’t make any difference in this case.