Java Pattern RegEx for Conditional Removal of Whitespaces

Question

I have searched for an answer for a couple of hours now and still nothing comes close to solving a specific programming predicament. This is neither for school or work. I'm developing an app that needs to perform pre-defined data cleansing tasks based on regular expressions. One specific expression that I'm having issues with is that of removing whitespace characters between a word and a number. Below are example requirements:

word 123           ==> word123
123 word           ==> 123word
world 123 wide     ==> word123wide
world wide 123     ==> world wide123
world wide 123 456 ==> world wide123 456

RegEx lookaround seems to be the right approach but still can't figure out how to apply the expression for phrases having more than 2 word blocks.

Thanks in advance.

Answer 1

Use a combination of lookarounds and alternance between two Patterns, as such:

//                | preceded by digit
//                |      | one whitespace
//                |      |   | followed by non-digit
//                |      |   |      | OR
//                |      |   |      | | preceded by non-digit
//                |      |   |      | |      | one whitespace
//                |      |   |      | |      |   | followed by digit
String pattern = "(?<=\\d)\\s(?=\\D)|(?<=\\D)\\s(?=\\d)";
// test Strings
String test0 = "word 123";
String test1 = "123 word";
String test2 = "world 123 wide";
String test3 = "world wide 123";
String test4 = "world wide 123 456";
// testing output: replace all found matches
// (e.g. one per String in this case)
// with empty
System.out.println(test0.replaceAll(pattern, ""));
System.out.println(test1.replaceAll(pattern, ""));
System.out.println(test2.replaceAll(pattern, ""));
System.out.println(test3.replaceAll(pattern, ""));
System.out.println(test4.replaceAll(pattern, ""));

Output:

word123
123word
world123wide
world wide123
world wide123 456