Partition Key Vs Primary Key Vs Sort Key Vs Clustering Key

Primary Key

A normal SIMPLE primary key in the database

create table customer (
	id bigint PRIMARY KEY,
	data text
)

Partition Key, Clustering Key and Sort Key

A primary key can be COMPOSITE (COMPOUND) which means it includes multiple columns

create table customer_order (
	customer_id bigint,
	order_id bigint
	created_date timestamp,
	PRIMARY KEY(customer_id, order_id)
)

In this case, the primary key is both customer_id and order_id. As a result:

• customer_id: partition key
  • How the data is being distributed across the nodes
• order_id: sort key, clustering key
  • How the data is being sorted within a partition

Composite / compound key is just any regular key.

Note

We can have multiple clustering key or partition key, for example:

create table customer_order (
	customer_id bigint,
	account_id bigInt,
	order_id bigint
	created_date timestamp,
	data text,
	PRIMARY KEY((customer_id, account_id), created_date, order_id)
)

In this case,

• Partition key: customer_id, account_id
• Clustering key / Sort key: created_date, order_id