最近一段时间作东西,总是因为要处理可能不通字符集的文件而烦心,虽然遭到一个程序库,好象是模仿Mozilla的一个C库使用java实现的,但是普通的情况下使用这个库是不是太大了点.于是将GBK作为默认编码,主要判断utf-8,我实现了一个程序,在Windows下边如果使用记事本保存的TXT文件带有BOM当然很方便,但是如果没有BOM的话,就只有通过文件中的字符的编码来判断了,经过几次简单的实验,发现尚且可用。JAVA源码如下,恳请指正:
- static String get_charset(File file) {
- String charset = "GBK";
- byte [] first3Bytes = new byte[3];
- try {
- boolean checked = false;
- BufferedInputStream bis = new BufferedInputStream(new FileInputStream(file));
- bis.mark(0);
- int read = bis.read(first3Bytes, 0, 3);
- if (read == -1) return charset;
- if (first3Bytes[0] == (byte)0xFF && first3Bytes[1] == (byte)0xFE) {
- charset = "UTF-16LE";
- checked = true;
- }
- else if(first3Bytes[0] == (byte)0xFE && first3Bytes[1] == (byte)0xFF) {
- charset = "UTF-16BE";
- checked = true;
- }
- else if(first3Bytes[0] == (byte)0xEF && first3Bytes[1] == (byte)0xBB && first3Bytes[2] == (byte)0xBF) {
- charset = "UTF-8";
- checked = true;
- }
- bis.reset();
- if (!checked) {
- int len = 0;
- int loc = 0;
- while ((read = bis.read()) != -1) {
- loc ++;
- if (read >= 0xF0)
- break;
- if (0x80<=read && read <= 0xBF) //单独出现BF以下的,也算是GBK
- break;
- if (0xC0<=read && read <= 0xDF) {
- read = bis.read();
- if (0x80<= read && read <= 0xBF)//双字节 (0xC0 - 0xDF) (0x80 - 0xBF),也可能在GB编码内
- continue;
- else
- break;
- } else if (0xE0 <= read && read <= 0xEF) {//也有可能出错,但是几率较小
- read = bis.read();
- if (0x80<= read && read <= 0xBF) {
- read = bis.read();
- if (0x80<= read && read <= 0xBF) {
- charset = "UTF-8";
- break;
- } else
- break;
- } else
- break;
- }
- }
- System.out.println(loc + " " + Integer.toHexString(read));
- }
- bis.close();
- } catch (Exception e) {
- e.printStackTrace();
- }
- return charset;
- }