Mastering Redis: Key Design and Data Type Use Cases

Redis Overview

Redis stores all data in memory, providing fast read/write operations. It supports rich data types—string, hash, list, set, sorted set, bitmap, hyperloglog—and persistence via AOF and RDB. All commands are atomic, and transactions are available.

Key Design

Use a colon‑separated namespace to encode hierarchical meaning: table:primaryKeyName:primaryKeyValue:field. Example for caching a user's email:

SET user:id:1:email [email protected]

String

Binary‑safe key‑value pairs. Suitable for storing single fields, JSON blobs, or binary data. Common operations:

Cache a column value (e.g., email).

Store entire objects as JSON (though hashes are often more efficient).

Auto‑increment counters with INCR when the value is numeric.

Hash

Represents a map of field‑value pairs under a single key, analogous to a relational table row. Example: HSET user:1 name zj email [email protected] Hashes reduce memory usage and avoid JSON parsing overhead when storing objects.

List

Ordered collection implemented as a doubly linked list. Insertion at head ( LPUSH) and removal at tail ( RPOP) are O(1). Typical uses:

Message queues.

Fetching the most recent items (e.g., news feed).

Set

Unordered collection of unique elements with O(1) add, remove, and membership checks. Supports set operations (intersection, union, difference). Example use cases:

Find mutual friends: SINTER user:wade user:james.

Identify common tags or interests across entities.

Sorted Set

Each element is associated with a score, automatically ordered by that score. Common commands: ZADD – add element with score. ZREVRANGE – retrieve elements in descending score order. ZINCRBY – increment an element's score.

Typical scenarios:

Rank friends by intimacy score.

Sort articles by view count or likes.

Practical Examples

Storing a field by primary key: SET user:id:1:email [email protected] Object storage with hash: HSET user:1 name zj email [email protected] Auto‑increment ID: INCR user:id:counter Message queue with list:

LPUSH queue:tasks "task1"
RPOP queue:tasks

Mutual friends using set intersection:

SADD user:wade james melo paul kobe
SADD user:james wade melo paul kobe
SINTER user:wade user:james

Friend ranking with sorted set:

ZADD user:kobe 80 james 90 wade 85 melo 90 paul
ZREVRANGE user:kobe 0 -1 WITHSCORES
ZINCRBY user:kobe 15 james

Choosing the appropriate Redis data type and designing clear, hierarchical keys maximizes performance, reduces memory consumption, and simplifies data access patterns.