你可以完全避免缓冲。

从http://www.exampledepot.com/egs/java.io/File2ByteArray.html复制的代码(是的，它非常冗长，但需要的内存大小是其他解决方案的一半)。

// Returns the contents of the file in a byte array.
public static byte[] getBytesFromFile(File file) throws IOException {
    InputStream is = new FileInputStream(file);

    // Get the size of the file
    long length = file.length();

    // You cannot create an array using a long type.
    // It needs to be an int type.
    // Before converting to an int type, check
    // to ensure that file is not larger than Integer.MAX_VALUE.
    if (length > Integer.MAX_VALUE) {
        // File is too large
    }

    // Create the byte array to hold the data
    byte[] bytes = new byte[(int)length];

    // Read in the bytes
    int offset = 0;
    int numRead = 0;
    while (offset < bytes.length
           && (numRead=is.read(bytes, offset, bytes.length-offset)) >= 0) {
        offset += numRead;
    }

    // Ensure all the bytes have been read in
    if (offset < bytes.length) {
        throw new IOException("Could not completely read file "+file.getName());
    }

    // Close the input stream and return bytes
    is.close();
    return bytes;
}

20年后，终于有了一个不需要第三方库的简单解决方案，这要感谢Java 9:

InputStream is;
…
byte[] array = is.readAllBytes();

还要注意方便的方法readNBytes(byte[] b, int off, int len)和transferTo(OutputStream)来解决重复的需求。

如果由于某种原因，它不在表中，则将其包装在DataInputStream中，只需使用read来锤击它，直到它给您一个-1或您所要求的整个块。

public int readFully(InputStream in, byte[] data) throws IOException {
    int offset = 0;
    int bytesRead;
    boolean read = false;
    while ((bytesRead = in.read(data, offset, data.length - offset)) != -1) {
        read = true;
        offset += bytesRead;
        if (offset >= data.length) {
            break;
        }
    }
    return (read) ? offset : -1;
}

下面是一个优化的版本，尽量避免复制数据字节:

private static byte[] loadStream (InputStream stream) throws IOException {
   int available = stream.available();
   int expectedSize = available > 0 ? available : -1;
   return loadStream(stream, expectedSize);
}

private static byte[] loadStream (InputStream stream, int expectedSize) throws IOException {
   int basicBufferSize = 0x4000;
   int initialBufferSize = (expectedSize >= 0) ? expectedSize : basicBufferSize;
   byte[] buf = new byte[initialBufferSize];
   int pos = 0;
   while (true) {
      if (pos == buf.length) {
         int readAhead = -1;
         if (pos == expectedSize) {
            readAhead = stream.read();       // test whether EOF is at expectedSize
            if (readAhead == -1) {
               return buf;
            }
         }
         int newBufferSize = Math.max(2 * buf.length, basicBufferSize);
         buf = Arrays.copyOf(buf, newBufferSize);
         if (readAhead != -1) {
            buf[pos++] = (byte)readAhead;
         }
      }
      int len = stream.read(buf, pos, buf.length - pos);
      if (len < 0) {
         return Arrays.copyOf(buf, pos);
      }
      pos += len;
   }
}

你可以试试仙人掌:

byte[] array = new BytesOf(stream).bytes();

您可以使用Apache Commons IO来处理这个任务和类似的任务。

IOUtils类型有一个静态方法来读取InputStream并返回一个字节[]。

InputStream is;
byte[] bytes = IOUtils.toByteArray(is);

这将在内部创建一个ByteArrayOutputStream并将字节复制到输出，然后调用toByteArray()。它通过以4KiB为块复制字节来处理大文件。