java - String.replaceAll to replace all characters not in pattern

Question

I have a Java regex:

^[a-zA-Z_][a-zA-Z0-9_]{1,126}$

It means:

• Begin with an alphabetic character or underscore character.
• Subsequent characters may include letters, digits or underscores.
• Be between 1 and 127 characters in length.

Now, I want to replace a string having characters not in that regex with a underscore.

Example:

final String label = "23_fgh99@#";
System.out.println(label.replaceAll("^[^a-zA-Z_][^a-zA-Z0-9_]{1,126}$", "_"));

But the result is still 23_fgh99@# .

How can I "convert" it to _3_fgh99__ ?

Answer 1

Use this code :

final String label = "23_fgh99@#";
System.out.println(label.replaceAll("^[^a-zA-Z_]|(?<!^)[^a-zA-Z0-9_]", "_"));

It outputs _3_fgh99__ .

To remove what is "not in the original pattern", you need to negate the first character class and only check a character at the beginning ( ^[^a-zA-Z_] ), and then check other characters not at the beginning with the negated second character class ( (?<!^)[^a-zA-Z0-9_] ). Then, we just use an alternation symbol | to apply both patterns in 1 replacement operation.