Split on capitalized words not between underscores

Given the following string: ThisIsA_SimpleTest_Case

I want to split on all capitalized words not between underscores and on the first underscore of a string between underscores.

The expected splitted result: This Is A SimpleTest Case

I came up with the following none working regex, for the Java regex flavor:

(?=_[a-zA-Z]*_|[A-Z])

But this ofcourse doesn't work since it's an or and not an and. Also this splits on all capitalized words within underscores which is something I want to ignore.

Wiktor is right, it should be easier to try to match instead of splitting on what you don't want.

But because it's a fun challenge, I got one that will split it like you wanted.
_|(?<!_)(?=[A-Z])(?=[^_]*(?:_[^_]*_[^_]*)*[^_]*$)

Also works with multiple pairs of underscores.
(It can certainly be improved, I might try to simplify it)

The idea is :

• _| Split on any underscore removing it from the final list.
• (?<!_) Not right after an underscore. If you don't do that, you might get empty matches after the split (cases already handled by the _|). Can be skipped if you don't care.
• (?=[A-Z]) Split before capital letters.
• (?=[^_]*(?:_[^_]*_[^_]*)*[^_]*$) But it must be followed by an even number of underscores. If there are an odd number, it means you're between 2 and it should not split. I assume there can't be an odd number of underscores in the string.

Test at https://regex101.com/r/Iov1Yl/1/

You might split on:

(?=(?<!_)[A-Z](?![A-Za-z]*_))|(?<!_[A-Za-z]{0,1000}|^)(?=[A-Z])|_

• (?=(?<!_)[A-Z](?![A-Za-z]*_)) If it is a position where a char A-Z is not directly preceded by _ and has no _ at the right
• | Or
• (?<!_[A-Za-z]{0,1000}|^)(?=[A-Z]) If it is a position where what is at the left is not an underscore or the start of the string, and what is directly at the right is a char A-Z
• | Or
• _ Match an underscore

Regex demo | Java demo

Example code

String regex = "(?=(?<!_)[A-Z](?![A-Za-z]*_))|(?<!_[A-Za-z]{0,1000}|^)(?=[A-Z])|_";
String str = "ThisIsA_SimpleTest_Case";
String[] parts = str.split(regex);

for (String part : parts)
	System.out.println(part);

Output

This
Is
A
SimpleTest
Case

Another approach before split:

The string is changed before split, see context:

public static void main(String[] args) {
    String input = "ThisIsA_SimpleTest_Case";
    String inputReplace1 =  input.replaceAll("_(\\w+[a-z])([A-Z]\\w+)_", ",$1#$2");
    String inputReplace2 = inputReplace1.replaceAll("(?<=[a-z])(?=[A-Z])", ",");
    String inputReplace3 = inputReplace2.replaceAll("#", "");
    System.out.println(Arrays.asList(inputReplace3.split(",")));
}

Output:

[This, Is, A, SimpleTest, Case]