从 Java 中的字符串中删除不可打印的字符

Question

I have string as below:

String s = "$$$$A very beautiful girl having loads of £££££ in her 20�s.";

I went through some StackOverflow responses and tried below:

s.replaceAll("[^\\x00-\\x7F]", " ");

s.replaceAll("[^\\p{ASCII}]", " ");

Both of them are removing the wierd question mark , but they are also removing the pound(£) sign retaining the dollar($) sign. I need to retain the currency symbol. Can you suggest a more suitable approach?

Also, is there any other library to do this rather than using the regex?

Answer 1

Try using:

s.replaceAll("[^\\x00-\\xFF]", " ");

Your problem is, pound sign is a part of Latin-1 Supplement Unicode block, which is not included when you filter upto 7F .

Answer 2

To efficiently remove all unprintable characters from a string, including the often-overlooked Unicode control codes that have been exploited by hackers:

    String broken = "\r\nhello world\b\u200E\uDB80";

    StringBuilder fixed = broken.codePoints()
        .filter(c -> {
          switch (Character.getType(c)) {
            case Character.CONTROL:
            case Character.FORMAT:
            case Character.PRIVATE_USE:
            case Character.SURROGATE:
            case Character.UNASSIGNED:
              return false;
            default:
              return true;
          }
        })
        .collect(StringBuilder::new, StringBuilder::appendCodePoint, StringBuilder::append);

    assertEquals("hello world", fixed.toString());

If you want to remove other character classes then simply include them in the case statement. This implements a blacklist . If you prefer a whitelist then you can invert the logic to return true when a character is an acceptable type and return false for all others.