用StringBuilder和StringBuffer实现的Unicode解码方法的比较（Java）

初衷是用正则来写一个Unicode字符串转码的方法，一开始是打算结合StringBuilder写的，但是看到jdk7的Matcher.appendReplacement文档中一段示例代码用了Matcher.appendReplacement，原来已经有专门做替换用的方法了。

 Pattern p = Pattern.compile("cat");
 Matcher m = p.matcher("one cat two cats in the yard");
 StringBuffer sb = new StringBuffer();
 while (m.find()) {
     m.appendReplacement(sb, "dog");
 }
 m.appendTail(sb);
 System.out.println(sb.toString());

但是印象中StringBuilder性能应该更好啊，因为StringBuffer用synchronized实现的，所以写了简单测试测了一下两种实现（测试环境是JDK7）：

    @Before
    public void before(){
        for (int i = 0; i < 100; i++) {
            assertEquals(excepted, unicode2StringWithStringBuffer(input));
            assertEquals(excepted, unicode2StringWithStringBuilder(input));
        }
    }

    @Test
    public void testUnicode2StringWithStringBuilder() {
        long start = System.currentTimeMillis();
        for (int i = 0; i < COUNT; i++) {
            unicode2StringWithStringBuilder(input);
        }
        System.out.println(String.format("v1 StringBuilder %s takes: %s", COUNT, (System.currentTimeMillis() - start)));
    }

    @Test
    public void testUnicode2StringWithStringBuffer() {
        long start = System.currentTimeMillis();
        for (int i = 0; i < COUNT; i++) {
            unicode2StringWithStringBuffer(input);
        }
        System.out.println(String.format("v2 StringBuffer  %s takes: %s", COUNT, (System.currentTimeMillis() - start)));
    }

    private static final int COUNT = 10000000;
    private static final String excepted = "请求失败，参数错误:[action]";
    private static final String input = "\u8bf7\u6c42\u5931\u8d25\uff0c\u53c2\u6570\u9519\u8bef:[action]";
    private static final Pattern patternUnicode = Pattern.compile("\\\\u([0-9a-zA-Z]{4})");

    private static String unicode2StringWithStringBuilder(final String unicode) {
        if (unicode != null) {
            try {
                Matcher matcher = patternUnicode.matcher(unicode);
                StringBuilder stringBuilder = new StringBuilder(unicode);
                int offset = 0; //StringBuilder替换长度不等的字符产生的位置偏移
                while (matcher.find()) {
                    String current = matcher.group();
                    String code = matcher.group(1);
                    String ch = String.valueOf((char) Integer.parseInt(code, 16));
                    stringBuilder.replace(matcher.start() + offset, matcher.end() + offset, ch);

                    offset += 1 - current.length(); //1为ch长度
                }
                return stringBuilder.toString();
            } catch (Exception e) {
                e.printStackTrace();
                return unicode;
            }
        } else {
            return unicode;
        }
    }

    private static String unicode2StringWithStringBuffer(final String unicode) {
        if (unicode != null) {
            try {
                Matcher matcher = patternUnicode.matcher(unicode);
                StringBuffer sb = new StringBuffer();
                while (matcher.find()) {
                    matcher.appendReplacement(sb, String.valueOf((char) Integer.parseInt(matcher.group(1), 16)));
                }
                matcher.appendTail(sb);
                return sb.toString();
            } catch (Exception e) {
                e.printStackTrace();
                return unicode;
            }
        } else {
            return unicode;
        }
    }

1亿次和1千万次的执行结果分别是：

v2 StringBuffer  1000000 takes: 1815
v1 StringBuilder 1000000 takes: 1364

v2 StringBuffer  10000000 takes: 14107
v1 StringBuilder 10000000 takes: 13316

不知道这个测试科不科学，结果确实是StringBuilder快一些。

转载于:https://www.cnblogs.com/liqipeng/p/7875399.html