Java检测文件是否UTF8编码

介绍UTF-8编码规则

1. UTF-8 编码字符理论上可以最多到 6 个字节长, 然而 16 位 BMP 字符最多只用到 3 字节长. Bigendian UCS-4 字节串的排列顺序是预定的.

2. 字节 0xFE 和 0xFF 在 UTF-8 编码中从未用到.
3. 下列字节串用来表示一个字符. 用到哪个串取决于该字符在 Unicode 中的序号.
4. U-00000000 - U-0000007F: 0xxxxxxx
5. U-00000080 - U-000007FF: 110xxxxx 10xxxxxx
6. U-00000800 - U-0000FFFF: 1110xxxx 10xxxxxx 10xxxxxx
7. U-00010000 - U-001FFFFF: 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
8. U-00200000 - U-03FFFFFF: 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
9. U-04000000 - U-7FFFFFFF: 1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
11. xxx 的位置由字符编码数的二进制表示的位填入. 越靠右的 x 具有越少的特殊意义. 只用最短的那个足够表达一个字符编码数的多字节串. 注意在多字节串中, 第一个字节的开头"1"的数目就是整个串中字节的数目.

12. 例如: Unicode 字符 U+00A9 = 1010 1001 (版权符号) 在 UTF-8 里的编码为:
13. 11000010 10101001 = 0xC2 0xA9
14. 而字符 U+2260 = 0010 0010 0110 0000 (不等于) 编码为:
15. 11100010 10001001 10100000 = 0xE2 0x89 0xA0
17. 特殊规则: 文件头三个字节用16进制表示是EFBBBF, 此规则不通用, 由编辑工具定义.

19. 这种编码的官方名字拼写为 UTF-8, 其中 UTF 代表 UCS Transformation Format. 请勿在任何文档中用其他名字 (比如 utf8 或 UTF_8) 来表示 UTF-8, 当然除非你指的是一个变量名而不是这种编码本身.

复制代码

源码实现:

1. package com.yy.game.test;

3. import java.io.BufferedInputStream;

4. import java.io.File;
5. import java.io.FileInputStream;
6. import java.io.IOException;
7. import java.io.InputStream;
8. import java.nio.CharBuffer;
9. import java.nio.MappedByteBuffer;
10. import java.nio.channels.FileChannel;
11. import java.nio.channels.FileChannel.MapMode;
12. import java.nio.charset.Charset;
13. import java.nio.charset.CharsetDecoder;
14. import java.nio.charset.CoderResult;
16. public class UTF8Checker {

18. public static void main(String[] args) throws IOException {

19. File dir = new File("F:\\test");
20. for (File file : dir.listFiles()) {
21. System.out.format("%s: %s, %s%n", file, check(file), check2(file));
22. }
23. }
25. /**

26. * JDK自带API实现
27. */
28. @SuppressWarnings("resource")
29. public static boolean check2(File file) throws IOException {
30. long start = System.nanoTime();
31. FileChannel fc = null;
32. try {
34. fc = new FileInputStream(file).getChannel();

36. MappedByteBuffer buf = fc.map(MapMode.READ_ONLY, 0, fc.size());

38. Charset utf8 = Charset.forName("UTF-8");

39. CharsetDecoder decoder = utf8.newDecoder();
40. CharBuffer cbuf = CharBuffer.allocate((int) (buf.limit() * decoder.averageCharsPerByte()));
41. CoderResult result = decoder.decode(buf, cbuf, true);
42. return !result.isError();
43. } finally {
44. if (fc != null) {
45. fc.close();
46. }
47. long end = System.nanoTime();
48. System.out.println("used(ns):" + (end - start));
49. }
51. }

53. /**

54. * 自定义实现
55. */
56. public static boolean check(File file) throws IOException {
57. long start = System.nanoTime();
58. InputStream in = null;
59. try {
60. in = new BufferedInputStream(new FileInputStream(file));
61. StreamBuffer sbuf = new StreamBuffer(in, 1024);
62. if (sbuf.next() == 0xEF && sbuf.next() == 0xBB && sbuf.next() == 0xBF) {
63. return true;
64. }
65. sbuf.redo();
66. // 1. U-00000000 - U-0000007F: 0xxxxxxx
67. // 2. U-00000080 - U-000007FF: 110xxxxx 10xxxxxx
68. // 3. U-00000800 - U-0000FFFF: 1110xxxx 10xxxxxx 10xxxxxx
69. // 4. U-00010000 - U-001FFFFF: 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
70. // 5. U-00200000 - U-03FFFFFF: 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
71. // 6. U-04000000 - U-7FFFFFFF: 1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
72. for (int ch = 0; (ch = sbuf.next()) != -1;) {
73. int n = 0;
74. if (ch <= 0x7F) {
75. n = 1;
76. } else if (ch <= 0xBF) {
77. return false;
78. } else if (ch <= 0xDF) {
79. n = 2;
80. } else if (ch <= 0xEF) {
81. n = 3;
82. } else if (ch <= 0xF7) {
83. n = 4;
84. } else if (ch <= 0xFB) {
85. n = 5;
86. } else if (ch <= 0xFD) {
87. n = 6;
88. } else {
89. return false;
90. }
91. while (--n > 0) {
92. if ((sbuf.next() & 0x80) != 0x80) {
93. return false;
94. }
95. }
96. }
97. return true;
98. } finally {
99. if (in != null) {
100. in.close();
101. }
102. long end = System.nanoTime();
103. System.out.println("used(ns):" + (end - start));
104. }
106. }

108. static class StreamBuffer {

110. final InputStream in;

111. final byte[] buf;
112. int pos = -1;// 初始值为-1,表示指针尚未移动.
113. int len;
115. public StreamBuffer(InputStream in, int size) {

116. this.in = in;
117. if (size < 3) {
118. size = 3;
119. }
120. this.buf = new byte[size];
121. }
123. public void redo() {

124. this.pos = 0;
125. }
127. public int next() throws IOException {

128. if (len > 0 || pos < 0) {
129. if (++pos == len) {
130. if ((len = in.read(buf)) == 0) {
131. return -1;
132. }
133. pos = 0;
134. }
135. return this.buf[this.pos] & 0xFF;
136. } else {
137. return -1;
138. }
139. }
141. }

143. }

复制代码

在本机测试, JDK原生API需要创建CharBuffer,性能明显慢了25%以上.

1. used(ns):472420

2. used(ns):4490075
3. F:\test\b334d5fd-b8a7-48f4-9099-f6011c7e5a48.sql: true, true
4. used(ns):122515
5. used(ns):343490
6. F:\test\b334d5fd-b8a7-48f4-9099-f6011c7e5a482.sql: false, false
7. used(ns):55164
8. used(ns):82425
9. F:\test\test.sql: false, false

复制代码