Split string with alternative comma (,)

Question

I know how to tokenize the String, but the Problem is I want to tokenize the as shown below.

String st = "'test1, test2','test3, test4'";

What I've tried is as below:

st.split(",");

This is giving me output as:

'test1
test2'
'test3
test4'

But I want output as:

'test1, test2'
'test3, test4'

How do i do this?

Answer 1

Since single quotes are not mandatory , split will not work, because Java's regex engine does not allow variable-length lookbehind expressions. Here is a simple solution that uses regex to match the content, not the delimiters:

String st = "'test1, test2','test3, test4',test5,'test6, test7',test8";
Pattern p = Pattern.compile("('[^']*'|[^,]*)(?:,?)");
Matcher m = p.matcher(st);
while (m.find()) {
    System.out.println(m.group(1));
}

Demo on ideone .

You can add syntax for escaping single quotes by altering the "content" portion of the quoted substring (currently, it's [^']* , meaning "anything except a single quote repeated zero or more times).

Answer 2

The easiest and reliable solution would be to use a CSV parser. Maybe Commons CSV would help.

It will scape the strings based on CSV rules. So even '' could be used within the value without breaking it.

A sample code would be like: ByteArrayInputStream baos = new ByteArrayInputStream("'test1, test2','test3, test4'".getBytes());

    CSVReader reader = new CSVReader(new InputStreamReader(baos), ',', '\'');

    String[] read = reader.readNext();
    System.out.println("0: " + read[0]);
    System.out.println("1: " + read[1]);

    reader.close();

This would print:

0: test1, test2
1: test3, test4

If you use maven you can just import the dependency:

    <dependency>
        <groupId>net.sf.opencsv</groupId>
        <artifactId>opencsv</artifactId>
        <version>2.0</version>
    </dependency>

And start using it.