从二进制文件Java错误读取int

Question

I'm trying to read a date (set of 6 integers) and temperature (double) from binary .dat file.

After multiple tries I finally got to the stage where the file is working, but it's returning int in the format I cannot recognize. Eg. date 2017-03-02 11:33 , and temperature 3.8 is read as:

Measure : 515840-1024-1024 2816 8512 241591910 temperature: 1.9034657819129845E185

Any ideas, how to change the code?

public void readFile() {

                try {
                    DataInputStream dis = null;
                    BufferedInputStream bis = null;
                    try {
                        FileInputStream fis = new FileInputStream(fileLocation);

                        int b;
                        bis = new BufferedInputStream(fis);
                        dis = new DataInputStream(fis);

                        while ((b = dis.read()) != -1) {

    System.out.println("Measure : " + dis.readInt() + "-" 
    + dis.readInt() + "-" + dis.readInt() + " " + 
    dis.readInt() + " " + dis.readInt() + " "
    + dis.readInt() + " Temperature: "+ dis.readDouble());

                        }
                    } finally {
                        dis.close();
                    }
                } catch (FileNotFoundException ex) {
                    ex.printStackTrace();
                } catch (EOFException f) {
                    f.printStackTrace(); 
                } catch (IOException e) {
                    e.printStackTrace();
                }

            } // readFile   

Answer 1

while ((b = dis.read()) != -1) {

The problem is here. This reads and discards a byte of the file on every iteration, so all subsequent reads are out of sync.

The correct way to loop with a DataInputStream or ObjectInputStream is with a while (true) loop, and terminating it when read() returns -1, readLine() returns null , or readXXX() for any other X throws EOFException.

Note that you don't normally need to log or print a stack trace on EOFException , as it's a normal loop termination condition ... unless you had reason to expect more data, eg your file started with a record count that you haven't reached yet, which might indicate that the file was truncated and therefore corrupt.