How to remove non-valid unicode characters from strings in java

Question

I am using the CoreNLP Neural Network Dependency Parser to parse some social media content. Unfortunately, the file contains characters which are, according to fileformat.info,

Answer 1

Observed the negative impact in other places when we do replaceAll. So, I propose to replace characters if it is non BPM characters like below

private String removeNonBMPCharacters(final String input) {
    StringBuilder strBuilder = new StringBuilder();
    input.codePoints().forEach((i) -> {
        if (Character.isSupplementaryCodePoint(i)) {
            strBuilder.append("?");
        } else {
            strBuilder.append(Character.toChars(i));
        }
    });
    return strBuilder.toString();
}