Variable Length Encoding

In general, when we declare a type as integer, then it must be 32 bits or 64 bits. But in case of smaller numbers, we don't need so many bits at all to represent small numbers. This is where variable length encoding is used.

Used only for storing and transmission

This is a technique that's used only for storing to save disk space and during transmission to reduce payload space.

When the same data is loaded into the CPU registers, it must be padded with additional zeros to fit into the register. This is the task of the application.

This is only an encoding technique.

How it's achieved?

1. Integers (VARINT) - First bit of the byte is used to signal if it's the last byte. If it's 1, it means it's not the last byte and it must read the next byte as well. It reads through the entire data like this and forms the actual value with it.

2. Strings (VARCHAR) - The initial bits store the length of the string, followed by the actual string.