Create a case-insensitive regular expression from a string in Ruby

Question

Let's say that I have an arbitrary string like

`A man + a plan * a canal : Panama!`

and I want to do a regex search for strings that are the same other than case. That is, this regular expression should match the string

`a man + A PLAN * a canal : PaNaMa!`

I take it the best approach is to backslash-escape every character with a special meaning in Ruby regular expressions, and then do Regexp.new with that string and Regexp::IGNORECASE as arguments. Is that right? Is there a tried-and-true regular expression for converting arbitrary strings into literal regular expressions?

By the way, I ultimately want to use this regular expression to do an arbitrary case-insensitive MongoDB query. So if there's another way I could be doing that, please let me know.

Solution 1

You can use Regexp.escape to escape all the characters in the string that would otherwise be handled specially by the regexp engine.

Regexp.new(Regexp.escape("A man + a plan * a canal : Panama!"), Regexp::IGNORECASE)

or

Regexp.new(Regexp.escape("A man + a plan * a canal : Panama!"), "i")

Solution 2

Ruby regexes can interpolate expressions in the same way that strings do, using the #{} notation. However, you do have to escape any regex special characters. For example:

input_str = "A man + a plan * a canal : Panama!"
/#{Regexp.escape input_str}/i

Solution 3

If you know the regular expression you want already, you can add "i" after the expression (eg /the center cannot hold it is too late/i) to make it case insensitive.

Solution 4

A slightly more syntactic-sugary way to do this is to use the %r notation for Regexp literals:

input_str = "A man + a plan * a canal : Panama!"
%r(#{Regexp.escape(input_str)})i

Of course it comes down to personal preference.

View more solutions

Comments

[Trevor Burnham]

Right, but that's not the question I'm asking. I have an arbitrary string (from user input), not a regular expression. If the user enters a+b, for example, I want to be able to find A+b, a+B, or A+B, not aaaaab.

[Trevor Burnham]

Thanks, that was just what I was looking for! (Though as to MongoDB, I realized that if I'm doing this kind of search often, I should really store a downcased version of the string for performance reasons.)

[Steve Midgley]

This is such a great answer - it reads way better than the top voted response and feels more like Ruby to me to boot. Hope others will upvote this too..

[user2066657]

I can't unify what you wrote just above with your original question. Consider revising your question?

[cdpalmer]

I agree, I just upvoted this answer that now ties the accepted answer! One more vote!

[Ahmed Kooli]

Used this in a clash of code competition :D