A format says how many bytes each field takes and in what order. Packing writes values into those slots, and unpacking reads them back, which is how files and protocols are built.
About 1 min of reading
Pack and unpack a record.
Pack two short numbers.
Before you run it, say out loud what it will print. Then open the answer.
Read them back with the same layout.
Before you run it, say out loud what it will print. Then open the answer.
Show an int is four bytes.
Before you run it, say out loud what it will print. Then open the answer.
Ask for four bytes from two and fail.
Before you run it, say out loud what it will print. Then open the answer.
Read the first example one line at a time. A line that prints shows what appears on screen; the rest set things up quietly.
A binary file is a sequence of fields with known sizes. The format string records that layout, and both the writer and the reader use it.
If a format changes size, old files read wrong. Real formats put a version or a length at the front so a reader can tell which layout to use.
import struct
print(struct.pack("<HH", 1, 2))Two unsigned shorts, little-endian, make four bytes. The format string is the layout of the record.
import struct
print(struct.unpack("<HH", struct.pack("<HH", 1, 2)))Unpacking with the same layout returns the values. The layout is the contract between writer and reader.
import struct
print(len(struct.pack("<I", 1)))An unsigned int is always four bytes here. Fixed sizes are what let a reader know where each field starts.
import struct
print(struct.unpack("<I", b"\x01\x02"))Asking for four bytes from two fails. The layout has to match the bytes, which is the check that catches drift.
Each one is defined here on this site. Open the entry for the full definition and the other words it points at.