import java.text.MessageFormat; /* * @author majb */ public class StringCharTest { private static final int LENGTH = 4; /** * @param args */ public static void main(String[] args) { try{ String string = "測試"; byte[] bytes = string.getBytes(); System.out.println(bytes.length == LENGTH);//Expect true, but was false. System.out.println(new String(bytes)); int length = bytes.length; int i = 0; while(i < length) { System.out.print(MessageFormat.format("The bytes from {0} to {1} is :", i, i+2)); System.out.println(new String(bytes, i, 2)); i += 2; } bytes = null; bytes = string.getBytes("GB2312"); System.out.println(bytes.length);// The length is 6. System.out.println(new String(bytes, "GB2312")); }catch(Exception e){ // do nothing } } }

大家都知道字串getbytes方法如果沒有charsetName引數的話，是採用Java預設的Unicode編碼方式把字串轉化成位元組陣列，而Unicode編碼是採用2個位元組來表示每個字元。按照這個推理上面例子中"測試".getBytes().length就是4了，但是結果確實6。通過遍歷每個字元發現，好像在前面自動加了""。

而對於GB2312等其他編碼方式來說沒有上面的問題。