问题
What is the best way to achieve the matching position (the index that would be returned by =~) for each match when using gsub or scan?
回答1:
"hello".gsub(/./) { Regexp.last_match.offset(0).first }
=> "01234"
See Regexp.last_match and MatchData.
回答2:
I came to this problem from a different direction, and could not come up with a decent solution (that is, understandable, maintainable) to do this with either gsub or scan (both built-in methods of String class). So I asked "Why do it this way?..." and looked for more natural alternatives (thanks to Nash for pointing the general direction!):
#!/usr/bin/env ruby
# -*- encoding: utf-8 -*-
# capture_all_matches.rb
#
# Copyright © 2015 Lorin Ricker <Lorin@RickerNet.us>
#
# This program is free software, under the terms and conditions of the
# GNU General Public License published by the Free Software Foundation.
# See the file 'gpl' distributed within this project directory tree.
# This sample code demonstrates three ways to capture *all* of the offsets
# [begin,end,length] data for *all* matches scanned in a source string.
# Of course, each/any of the below examples could be turned into a class method
# of String &/or Regexp -- One wonders why these are not part of the built-in
# classes/methods?...
# An example source string (any will do):
s = "The fox hides in the box full of sox eating lox."
# 4^ 25^ 31^
# Use the literal pattern /f/ as an example --
# there are three "f"s in the sample source string;
# see indexes above...
p = /f/
# 1. Just report an array of the begin (start) position of each match:
mpos = []
m = i = 0
m = p.match( s, i ) { |k| j = k.begin(0); i = j + 1; mpos << j } while m
p mpos # => [4, 25, 31]
# 2. Make an array containing elements [begin,end] of matched substrings:
mpos = []
m = i = 0
m = p.match( s, i ) { |k| j = k.offset(0); i = j[0] + 1; mpos << j } while m
p mpos # => [[4, 5], [25, 26], [31, 32]]
# 3. Make an array containing elements [begin,end,length] of matched substrings:
mpos = []
m = i = 0
m = p.match( s, i ) { |k| j = k.offset(0); i = j[0] + 1;
j << j[1] - j[0]; mpos << j } while m
p mpos # => [[4, 5, 1], [25, 26, 1], [31, 32, 1]]
Demonstrate by pasting the the above into a Ruby source file (e.g., capture_all_matches.rb), then:
$ ruby capture_all_matches.rb
Note that the RegExp match method is (re)startable from an arbitrary offset in the source string, so it's just a matter of capturing that "last matched offset" and iterating from there.
Need just the starting-offset of each match, or the start-&-end, or the start-end-length? Roll your own resulting array of results.
Hope this helps.
来源：https://stackoverflow.com/questions/9544319/matching-position-in-gsub-or-scan