Apache commons lang. Stringutils.splitByWholeseparator doing weird stuff.

Question

• As I understand from the official documentation, null as a separator string should split on WhiteSpace, so this System.out.println(JSON.serialize(StringUtils.splitByWholeSeparator("ab de fg", null))); should produce [ "ab" , "de" , "fg"]

• However, what I don't understand is why is an empty string "", also splitting on whitespace. Following produces the same output.

  System.out.println(JSON.serialize(StringUtils.splitByWholeSeparator("ab de fg", "")));

• Official documentation at https://commons.apache.org/proper/commons-lang/javadocs/api-2.6/org/apache/commons/lang/StringUtils.html#splitByWholeSeparator%28java.lang.String,%20java.lang.String%29 doesn't mention any such effect.

Answer 1

You're right. There's a lack in the documentation. null and empty String separator will produce a split on whitespace.

Internally the code is doing this:

    if (separator == null || EMPTY.equals(separator)) {
        // Split on whitespace.
        return splitWorker(str, null, max, preserveAllTokens);
    }

So these two calls will produce the same result:

StringUtils.splitByWholeSeparator("ab de fg", null);

StringUtils.splitByWholeSeparator("ab de fg", "");