I will admit that I find regular expressions somewhat hard to parse. While the thinking process needed to write a regular expression is iterative and hence results in the correct expression suitable for the situation, once it is written it lacks readability. I have faced that pain of trying to grok a complex regular expression written years ago. So I was pleasantly surprised when I saw this example from Hacker News user WestCoastJustin. I found it very easy to understand because of the names he had assigned to each capture group in the expression.
Ruby
Ruby has supported named capture groups since version 1.9. Each match group is named by using ?<name> inside the match group and these matches are made available has a hash and each match can be accessed by using the match group name as the key.
WestCoastJustin’s example will look like this in Ruby:
#!/usr/bin/env ruby
test_string = "Today's date is: 9/28/2013."
match = test_string.match /(?<month>\d{1,2})\/(?<day>\d{1,2})\/(?<year>\d{4})/
puts match.inspect
puts match[:month]
puts match[:day]
puts match[:year]This is much easier to understand because looking at it, it makes it clear that the intent of matching the first group of 1 or 2 digits is to look for the month in the date. Similarly, the intent of the day and year matches are easier to understand.
This will result in:
> ruby /tmp/named_matches.rb
#<MatchData "9/28/2013" month:"9" day:"28" year:"2013">
9
28
2013##Python
Python’s re module supports named match groups using the ?P<name> pattern. After a pattern search, the results are placed in a dict.
#!/usr/bin/env ruby
import re
test_string = "Today's date is: 9/28/2013."
pattern = re.compile(r"(?P<month>\d{1,2})\/(?P<day>\d{1,2})\/(?P<year>\d{4})")
match_dict = pattern.search(test_string).groupdict()
print(match_dict)JavaScript does not support named capture groups. There are a lot of hacks aimed at providing this functionality. The XRegExp library supports named capture groups.