Struggling with regex code: CamelCase to camel_case

Question

I was able to transform the string MyClassName to my_class_name using a regex

However my solution did not work with MyOtherTClassName, that should transform to my_other_t_class_name.

Also, this didn't work on ClassNumber1 either, which should be transformed into class_number_1

Without getting into my solution, that was not good enough, I would like help with the regex code that transforms:

MyClassName -> my_class_name
MyOtherTClassName -> my_other_t_class_name
MyClassWith1Number -> my_class_with_1_number

Thanks,

Guy

Your samples are inconsistent (why is `CLass` treated as `Class`?), and you need to show your 'not good enough' attempts. — Niels Keurentjes, Jan 11 '14 at 14:55
@NielsKeurentjes i think MyOtherTCLassName was meant to be MyOtherTClassName , guy is this correct ? — eran otzap, Jan 11 '14 at 15:00
I thought that as well until I saw the same 'mistake' made twice. — Niels Keurentjes, Jan 11 '14 at 15:03
Why not post your existing solution to the first problem? If it solves the first case I can't imagine the next case(s) would be that hard to expand it to. — Chris, Jan 11 '14 at 15:05
Here's a post of someone doing this exact thing in Python: [Elegant Python function to convert CamelCase to camel_case?](http://stackoverflow.com/questions/1175208/elegant-python-function-to-convert-camelcase-to-camel-case) — valverij, Jan 11 '14 at 15:07
I'm partial to [my solution to a similar question](http://stackoverflow.com/a/7599674/433790). — ridgerunner, Jan 11 '14 at 17:43

KeyNone · Accepted Answer · 2014-01-11T15:31:09.887

The logic behind is that you want to convert every capital letter to its lower-case variant and preceed it (and every number) with an underscore.
For example a T becomes _t, 6 becomes _6.
The only exception is the very first character. You don't want to preceed it with an undersoce. The regex will handle this case with a negative lookbehind in order to not match the first character.

//using System.Text.RegularExpression

//your input
string input = "MyOtherTClass1Name";

//the regex
string result = Regex.Replace(
    input, 
    "((?<!^)[A-Z0-9])", //the regex, see below for explanation
    delegate(Match m) { return "_" + m.ToString().ToLower(); }, //replace function
    RegexOptions.None
);
result = result.ToLower(); //one more time ToLower(); for the first character of the input

Console.WriteLine(result);

For the regex itself:

(           #start of capturing group
  (?<!      #negative lookbehind
     ^      #beginning of the string
  )         #end of lookbehind
  [A-Z0-9]  #one of A-Z or 0-9
)           #end of capturing group

So we capture every capital letter and every number (except for the very first character) and replace them with a lower-case variant of themselves combined with a preceeding underscore.

What if there is a sequence of multiple upper case letters (like HTML?) You could add a negative lookahead as well — BlackBear, Jan 11 '14 at 15:46
As the OP wanted every occurence of (i.e) `A` gets replaced with `_a`. So in case of multiple upper case letters each and every of them gets replaced. — KeyNone, Jan 11 '14 at 15:56

score 0 · Answer 2 · answered May 17 '16 at 19:13

I recently had this problem and the previous answer works if there's only one digit but if there are two or more consecutive digits they each would have an underscore preceding them. I used this to convert it for me in PHP.

strtolower(preg_replace('/(?<!^)([A-Z])|(?<![0-9])([0-9])/', '_$1$2', $string))

The regex, I believe, should be the same as C# so I'll break that down.

(?<!       # negative look behind
   ^       # beginning of string
)
([A-Z])    # one of capital letters
|          # or
(?<!       # negative look behind
   [0-9]   # one of digits
)
([0-9])    # one of digits

The idea is the same for letters. Make sure it's not the beginning of the string. For the digits just make sure that the previous character is not also a digit. We don't have to worry about it being the beginning of a string because the string won't start with a digit.

Struggling with regex code: CamelCase to camel_case

2 Answers2