What is the difference between setting system property from command line and code?

Question

Here's the problem: I got some Chinese in my code, and I'm writing it into sqlite.

When I run my program in eclipse and read the string out of sqlite, it just works fine. But when I packaged the project to a jar and run it in command line. The Chinese string which is read out of sqlite is unreadable.

After some trials, I know that the problem lies in the file.encoding system property. When I run the jar using command: java -Dfile.encoding=UTF-8 -jar TK.jar, it works fine with Chinese word, But if I set the system property in code like: System.setProperty("file.encoding", "UTF-8");, it won't work.

So, What is the difference between setting the system property from command line and code? And could anyone tell me how to set file.encoding system property in code?

Thanks a lot!

To summary:

Remember to add charset when using String.getBytes() as well as new String() in order to avoid unreadable output from Chinese or Japanese when running program in different environment.

Answer 1

Somewhere in your code, you are probably relying on the default character set being UTF-8. For example, when you call String.getBytes() without specifying a character set, Java will use the default character set. If you always want UTF-8, then specify this when calling String.getBytes():

byte[] utf8bytes = text.getBytes("UTF-8");

Also, if you want to read a file with a specific character encoding, then specify the character encoding rather than relying on the default setting. For example:

InputStream is = new FileInputStream("utf8file.txt");
BufferedReader in = new BufferedReader(new InputStreamReader(is, "UTF-8"));

// Read text, file will be read as UTF-8
String line = in.readLine();

in.close();