二进制文件写入OutputStream后从InputStream读取不一致问题排查

Hey there! Let's figure out why your binary file is getting messed up after writing to an OutputStream and reading back from an InputStream. Binary files (like images, ZIPs, or executables) are super picky about byte-perfect handling—even tiny missteps can break them. Here are the most common issues to check:

1. You're using character-based streams instead of byte-based ones #

Character streams (like Reader/Writer) are designed for text, not binary data. They automatically convert bytes to characters using a default or specified encoding, which mangles raw binary values.

Wrong approach:

// Never use FileWriter for binary files!
FileWriter badWriter = new FileWriter("binary-file.bin");

Fix: Stick strictly to byte-based streams (InputStream/OutputStream and their subclasses like FileInputStream, BufferedOutputStream) for all binary operations.

2. Your read/write loop ignores actual bytes read #

A super common mistake is assuming inputStream.read(buffer) fills the entire buffer every time. This method returns the number of bytes actually read (or -1 when done). If you just write the full buffer every time, you'll end up appending leftover garbage bytes from previous reads to your output.

Wrong approach:

byte[] buffer = new byte[4096];
while (inputStream.read(buffer) != -1) {
    outputStream.write(buffer); // Writes the whole buffer, even if only part was filled
}

Fix: Use the returned byte count to write only the bytes you actually read:

byte[] buffer = new byte[4096]; // 4KB is a solid buffer size for most cases
int bytesRead;
while ((bytesRead = inputStream.read(buffer)) != -1) {
    outputStream.write(buffer, 0, bytesRead); // Only write the valid bytes
}

3. You're not flushing/closing streams properly #

Most OutputStream implementations (especially buffered ones) store data in an internal buffer before writing it to the destination. If you don't flush or close the stream, the final chunk of data might never get written.

Wrong approach:

OutputStream out = new BufferedOutputStream(new FileOutputStream("output.bin"));
// Write data...
// Forgot to close/flush—last bytes are stuck in the buffer!

Fix: Use try-with-resources to automatically close streams (this also flushes them):

try (InputStream in = new FileInputStream("original.bin");
     OutputStream out = new BufferedOutputStream(new FileOutputStream("copied.bin"))) {
    // Your read-write loop here
} catch (IOException e) {
    e.printStackTrace();
}

This ensures streams are properly closed even if an exception is thrown.

4. You're using specialized streams incorrectly #

Streams like DataInputStream/DataOutputStream are meant for reading/writing Java primitive types (ints, strings, etc.) in a standardized format—not raw binary files. Using methods like readUTF() or writeInt() on arbitrary binary data will corrupt the original bytes.

Fix: Only use these specialized streams if you're working with Java-specific data structures. For generic binary files, stick to basic byte streams.

5. Accidental data modification #

Double-check if you're modifying the binary data somewhere in your code before writing or after reading—like applying unnecessary compression, encoding, or filtering that alters the raw bytes.

If you can share your specific code snippet, we can pinpoint the exact issue, but these fixes should resolve most common binary file corruption cases with streams.

内容的提问来源于stack exchange，提问作者user7232539