字符串类拆分功能正在返回恶劣的结果

Question

below are 2 codes

System.out.println(",,,,,".split(",").length);

and

System.out.println(",,,,, ".split(",").length);

for first code the result it prints is 0 and for the second code result it prints is 6.

My question is why the split is not able to recognize "," when I am not adding extra space at the end but it is able to recognize it when I add extra space at the end.

Please note I have tried it with regex "\\\\s*,\\\\s" also but result is same.

Answer 1

I don't have a doc reference for this, but empirically what I saw in my testing of String#split is that if there are no actual text matches, then zero-width matches are also not returned in the array. So the following returns an empty array:

",,,,,".split(",")

However, if you add a space to the end of the series of commas and then do the same split, then there is a single matching space. As a result of this, the array comes back with all matches, including zero-width matches:

",,,,, ".split(",")

But, because there is no content in between the commas, I would interpret your real requirement as wanting to split each individual comma into a separate result. If so, then you can split using lookarounds, something like this:

String input = ",,,,,";
String[] parts = input.split("(?<=,)(?=,)");
for (String part : parts) {
    System.out.println(part);
}

This outputs:

,
,
,
,
,

Demo

Answer 2

split() in java by default removes trailing empty strings from result array. To keep empty, you can use split(delimiter, limit) with limit set to negative value, like this

System.out.println(",,,,," .split(",", -1).length);

Answer 3

Let's explore more see the interesting results of split below:

System.out.println(",,,,,,".split(",").length); // 0
System.out.println(",,,,,, ".split(",").length); // 7
System.out.println(",,, ,,,".split(",").length); // 4
System.out.println(" ,,,,,,".split(",").length); // 1

Wondering if why it's happening this is because below statement stated for the split method in docs:

Trailing empty strings are therefore not included in the resulting array.

Docs: https://docs.oracle.com/javase/7/docs/api/java/lang/String.html#split(java.lang.String)

if you don't want the split method to remove that spaces then you should use another split with limit:

public String[] split(String regex,int limit)

Docs: https://docs.oracle.com/javase/7/docs/api/java/lang/String.html#split(java.lang.String,%20int)

Example:

System.out.println(",,,,,,".split(",",-1).length); // 7
System.out.println(",,,,,, ".split(",",-1).length); // 7
System.out.println(",,, ,,,".split(",",-1).length); // 7
System.out.println(" ,,,,,,".split(",",-1).length); // 7

Answer 4

Forget documentation, I directly looked into the code and found the following piece of code in java.lang.String#split(java.lang.String, int) :-

while (resultSize > 0 && list.get(resultSize - 1).length() == 0) {
    resultSize--;
}

This proves that it is designed to remove the last element if it is empty . And keep doing it until the last element is not zero-length.

This feature is useful, for example, if you have a string a,b, it should return a and b in the resulting array and not the last blank character '' after the last comma.

If you do System.out.println(", ,,,".split(",").length); it will return 2 because the above while loop will keep decreasing the result from the right side until it finds something whose length is non-zero.

The above while loop is enclosed in if (limit == 0) . So if you want to count all, use a non-zero limit. If you don't want any limit, use a negative number like -1.