Packing unpacking values as bit fragments

suggest change

It is common for memory performance to compress multiple values into a single primitive value. This may be useful to pass various information into a single variable.

For example, one can pack 3 bytes - such as color code in RGB - into an single int.

Packing the values

// Raw bytes as input
byte[] b = {(byte)0x65, (byte)0xFF, (byte)0x31};

// Packed in big endian: x == 0x65FF31
int x = (b[0] & 0xFF) << 16  // Red
      | (b[1] & 0xFF) <<  8  // Green
      | (b[2] & 0xFF) <<  0; // Blue

// Packed in little endian: y == 0x31FF65
int y = (b[0] & 0xFF) <<  0
      | (b[1] & 0xFF) <<  8
      | (b[2] & 0xFF) << 16;

Unpacking the values

// Raw int32 as input
int x = 0x31FF65;

// Unpacked in big endian: {0x65, 0xFF, 0x31}
byte[] c = {
    (byte)(x >> 16),
    (byte)(x >>  8),
    (byte)(x & 0xFF)
};

// Unpacked in little endian: {0x31, 0xFF, 0x65}
byte[] d = {
    (byte)(x & 0xFF),
    (byte)(x >>  8),
    (byte)(x >> 16)
};

Found a mistake? Have a question or improvement idea? Let me know.

Table Of Contents