Why Does Java 8's String.split() Handle Empty Delimiters Differently Than Previous Versions?

Why in Java 8 Split Sometimes Removes Empty Strings at the Start of Result Array?

In Java 7 and prior, splitting a string using an empty delimiter ("") would result in an array containing both leading and trailing empty strings. This behavior changed in Java 8.

Java 8 Behavior

In Java 8, split on an empty delimiter no longer includes a leading empty string in the result array. However, it still removes trailing empty strings. This is due to a change in the implementation of String.split and Pattern.split.

Reference Implementation

Comparing the reference implementations of Pattern.split in Java 7 and Java 8, we find the following addition in Java 8:

if (index == 0 && index == m.start() && m.start() == m.end()) {
    // no empty leading substring included for zero-width match
    // at the beginning of the input char sequence.
    continue;
}

This code excludes zero-length matches at the beginning of the input string from the result array.

Maintaining Compatibility

Following Java 8 Behavior (Forward Compatibility)

To ensure consistent behavior across Java versions, modify your regular expressions as follows:

1. If the regex can match zero-length strings, add (?!A) at the end and wrap the original regex in a non-capturing group (?:...).
2. If the regex cannot match zero-length strings, no modification is necessary.
3. If unsure, perform both modifications.

Following Java 7 Behavior (Backward Compatibility)

There is no general solution for backward compatibility with Java 7 or prior. Custom implementations of split may be necessary.

The above is the detailed content of Why Does Java 8's String.split() Handle Empty Delimiters Differently Than Previous Versions?. For more information, please follow other related articles on the PHP Chinese website!