Ruby regular expression matching enumerator with named capture support

Question

Consider a string like this to extract the time information:

str = "Sun rises at 6:23 am & sets at 5:45 pm; Moon comes up by 7:20 pm and goes down by 3:45 am"

I wish to have an enumerator like scan but one that can get me the MatchData objects instead of arrays as available from scan.

For instance, I can write:

str.scan( /(?<time>\d:\d{2}) (?<meridiem>am|pm)/ ){ |arr| p arr }

to get:

["6:23", "am"] ["5:45", "pm"] ["7:20", "pm"] ["3:45", "am"]

But, I wonder if there something like this:

str.match_all( /(?<time>\d:\d{2}) (?<meridiem>am|pm)/ ){ |md| p md }

to get:

#<MatchData "6:23 am" time:"6:23" meridiem:"am"> #<MatchData "5:45 pm" time:"5:45" meridiem:"pm"> #<MatchData "7:20 pm" time:"7:20" meridiem:"pm"> #<MatchData "3:45 am" time:"3:45" meridiem:"am">

Saw an answer in a previous question, but I feel its an inelegant solution. So checking up in case things have changed over the last couple of years since the answer was posted.

Idden

The accepted answer on the question you referred seems short and quite elegant to me. You can implement this match_all based on that. – Margarethe

I agree with @Guilherme, but it's been helpful to me to think about your question, and see the earlier answer. – Surrey

The accepted answer to the linked question is as elegant as I can think of. If you are not satisfied with even that, then think of a more elegant answer by yourself. Considering your high standards, you should be able to do that by yourself. – Tentmaker

Thanks everyone; probably you are right! I somehow assumed such a match_all method should exist, assuming the requirement to be rather common. I must apologise in using the term "inelegant" for the quoted answer; I was just looking for a rather direct method instead. – Idden

Answer 1

Very identical to the answer you have already seen, but slightly different.

str = "Sun rises at 6:23 am & sets at 5:45 pm; Moon comes up by 7:20 pm ..."
str.gsub(/(?<time>\d:\d{2}) (?<meridiem>am|pm)/).map{ Regexp.last_match } 

#=> [#<MatchData "6:23 am" time:"6:23" meridiem:"am">, #<MatchData "5:45 pm" ...

Hen

I was thinking about a similar solution, but I didn't know Regexp.last_match existed. But somehow catching the MatchData from there seems problematic to me, when thinking about its scope. – Idden

Really wished something like match_all existed! Quite unbelievable why it doesn't. – Idden