Fundamantal Redis Commands

Redis provides essential commands for key-value operations. The most basic are:

• SET key value: Assigns a value to a key.
• GET key: Retrieves the value of a key. Returns nil if the key doesn’t exist.

To handle binary-safe strings correctly in the CLI, use redis-cli --raw.

Global Commands

• KEYS pattern: Matches keys using glob-style patterns (e.g., ?, *, [abc]). Avoid KEYS * in production—it’s O(n) and blocks the single-threaded server.

Generic Commands

• EXISTS key1 [key2 ...]: Returns the count of existing keys (O(1)).
• DEL key1 [key2 ...]: Deletes keys; returns the number deleted (O(1)).
• EXPIRE key seconds: Sets a TTL in seconds; returns 1 on success, 0 otherwise (O(1)).
• TTL key: Returns remaining TTL. -1 means no expiration; -2 means key doesn’t exist (O(1)).
• TYPE key: Returns the data type of the value stored at key.

String-Specific Commands

• APPEND key value: Appends to a string; creates key if missing. Returns new length.
• INCR key: Increments integer value by 1 (treats missing key as 0).
• DECR key: Decrements by 1.
• INCRBY key increment, DECRBY key decrement: Adjust by arbitrary integers.
• INCRBYFLOAT key increment: Supports floating-point adjustments.
• GETRANGE key start end: Extracts substring (inclusive, supports negative indices).
• SETRANGE key offset value: Overwrites part of a string; pads with \x00 if needed.
• STRLEN key: Returns byte length of string value.

Advanced SET options:

• SET key value [EX seconds | PX milliseconds] [NX | XX]
  • NX: Set only if key doesn’t exist.
  • XX: Set only if key exists.

Batch operations reduce network round trips:

• MSET key1 val1 key2 val2 ...
• MGET key1 key2 ...

Data Types and Internal Encodings

All Redis keys are strings, but values support multiple types with optimized internal representations:

• String: Encoded as int (for numeric strings), embstr (≤39 bytes), or raw (longer strings).
• Hash: Uses ziplist for small hashes; switches to hashtable when thresholds are exceeded.
• List: Historically used ziplist or linkedlist; since Redis 3.2, uses quicklist (a ziplist-backed linked list).
• Set: Encoded as intset (small sets of integers) or hashtable (larger or non-integer sets).
• Sorted Set (ZSet): Uses ziplist for small sets; skiplist for larger ones, enabling efficient range queries.

Use OBJECT ENCODING key to inspect the actual encoding.

Key Expiration Mechanism

Redis employs two complementary strategies:

1. Lazy Expiration: Checks expiration only when a key is accessed. If expired, it’s deleted before returning a result.
2. Periodic Expiration: Every 100ms, Redis samples a subset of expiring keys and deletes those that have past their TTL.

This hybrid approach balances CPU usage and memory efficiency without requiring full scans.

Why Redis Is Single-Threaded

Redis processes commands in a single thread to:

• Eliminate concurrency complexity (no locks or race conditions).
• Avoid CPU overhead from context switching.
• Leverage I/O multiplexing (e.g., epoll) to handle thousands of connections efficiently.
• Guarantee atomicity for individual commands.
• Maximize CPU cache locality for memory-bound operations.

However, long-running commands (e.g., KEYS *, large SMEMBERS) can block the entire server. Hence, O(n) operations on large datasets should be avoided.

Performance stems from in-memory storage, minimal command logic, and efficient event-driven I/O—not raw threading. For bulk deletion, FLUSHALL clears all database.