J2ee 断点续传 及 post请求中 包含 url参数及文件流

首先

client端：

HttpURLConnection conn = null;
BufferedInputStream fin = null;
BufferedOutputStream out = null;
URL reqUrl;
try {
reqUrl = new URL("http://<ip>:<port>/Emergency/phone/attachmentUpload");
conn = (HttpURLConnection) reqUrl.openConnection();
conn.setConnectTimeout(3000);
conn.setRequestMethod("PUT");
if(isCommit){
conn.setRequestProperty("isCommit", "true");
}else{
conn.setRequestProperty("isCommit", "false");
}
conn.setRequestMethod("PUT");
conn.setRequestProperty("Content-Type", "binary/octet-stream");
conn.setRequestProperty("offset", StringUtil.toString(offset));
conn.setDoOutput(true);
conn.setDoInput(true);
// 1M的chunk缓冲
conn.setChunkedStreamingMode(1024*1024);
out = new BufferedOutputStream(conn.getOutputStream());
fin = new BufferedInputStream(new FileInputStream(file));
byte[] buf = new byte[bufferSizeUpload];
int len = -1;
long currentUploadSize = offset;
fin.skip(offset);
while ((len = fin.read(buf)) != -1¤tUploadSize<offset+uploadSize) {
if(offset+uploadSize-currentUploadSize<bufferSizeUpload){
len = Integer.parseInt(StringUtil.toString(offset+uploadSize-currentUploadSize));
}
if(len>0){
if(out!=null){
out.write(buf, 0, len);
out.flush();
}
}
currentUploadSize += len;
}
} catch (SocketTimeoutException e) {
e.printStackTrace();
} catch(IOException e){
e.printStackTrace();
} finally {
try {
if (fin != null) {
fin.close();
fin = null;
}
if (out != null) {
out.close();
out = null;
}
if (conn != null) {
conn.disconnect();
conn = null;
}
} catch (IOException ioe) {
ioe.printStackTrace();
throw new RuntimeException("Release resource failed.");
}
}

如上所示：url参数通过setRequestProperty方法放在请求头中 ，而文件流放在 body 中。

同时需要注意 HttpURLConnection.setChunkedStreamingMode 此方法保证每次文件流达到指定大小就发送一次，避免了放在缓存并一次性传输中可能遇到的数据缺失。

http://blog.csdn.net/z69183787/article/details/8186918 可参考

服务端：

public String underUpload() throws IOException{
String data = request.getHeader("data");
String userInfo = request.getHeader("userInfo");
userInfo = new String(userInfo.getBytes("iso-8859-1"),"GBK");
//System.out.println(data);
//System.out.println(userInfo);
Map<String,String> data_map = gson.fromJson(data,attachVo.data_map.getClass());
if(data_map!=null) attachVo.data_map=data_map;
attachVo.user = gson.fromJson(userInfo,attachVo.user.getClass());
apiService.attachUpload(attachVo,new BufferedInputStream(request.getInputStream()));
//response.setStatus(200);
return null;
}

使用的是ssh架构 ，通过getHeader 得到 url参数，通过getInputStream得到文件流，同时还要控制好编码，以防中文乱码。

至于断点续传，基本思想是通过信息头的部分 传递每次传输的文件大小，与服务器端的文件大小匹配。并通过

RandomAccessFile raFile = new RandomAccessFile(uploadFile.getAbsolutePath(), "rw");
raFile.seek(uploadFile.length());

方法进行根据传输大小实时进行文件读写。

public static void saveFile(String destFilePathStr,String destFileName){
try {
File destFilePath = new File(destFilePathStr);
if(!destFilePath.exists()){
destFilePath.mkdirs();
destFilePath = null;
}
File destFile = new File(destFilePathStr+"//"+destFileName);
if(!destFile.exists()){
destFile.createNewFile();
}

} catch (Exception e) {
e.printStackTrace();
}
}

public static long uploadFile(String offset,String destFileName,BufferedInputStream bis)
throws IOException{
File uploadFile = new File(destFileName);
int len = 0;
byte[] bt = new byte[1024];
RandomAccessFile raFile = new RandomAccessFile(uploadFile.getAbsolutePath(), "rw");
raFile.seek(uploadFile.length());
while ((len = bis.read(bt)) > 0){
raFile.write(bt, 0, len);
}
long l = raFile.length();
try {
if(bis != null)
bis.close();
if (raFile != null)
raFile.close();

} catch (IOException e) {
l = 0;
e.printStackTrace();
}
return l ;
}

首先创建一个空文件，接下来根据每次接收的文件流及文件长度 写入文件。