Redis Command Generator

Redis is an in-memory data structure store used as a database, cache, message broker, and queue. Use Redis for: caching database results (sub-millisecond reads), session storage (fast auth checks), real-time leaderboards (sorted sets), pub/sub messaging (fan-out notifications), rate limiting (atomic counters), and task queues (RPUSH/BLPOP pattern).

SET key value EX 3600 is atomic — sets the value and TTL in one operation. Calling SET followed by EXPIRE is two operations and not atomic; a crash between them leaves a key with no TTL (persists forever). Always use SET ... EX/PX when creating keys with TTL. Use EXPIRE/PEXPIRE only to extend or add TTL to existing keys.

Use SET resource token NX PX ttl. NX=only set if not exists (atomic mutex). PX=TTL in milliseconds to auto-release if the holder crashes. Always store a unique token (UUID) and verify it before releasing — use a Lua script for atomic compare-and-delete. This is the core of the Redlock algorithm. Never use SETNX + EXPIRE (two separate commands) — it's not atomic.

LPOP/RPOP immediately return null if the list is empty — your worker must loop and poll. BLPOP/BRPOP block the connection until an item is available or the timeout expires. Blocking pops are more efficient for task queues — no polling loop needed. Use BLPOP for worker processes. Note: a blocking connection cannot be used for other commands while waiting.

Use HSET for sessions with multiple fields — HGET retrieves a single field without deserializing the whole session. SET requires serializing/deserializing a full JSON blob on every access. HSET is also more memory-efficient for many small fields. Use SET when the session is always accessed as a whole and rarely updated partially.

Strings: caching, counters, simple key-value. Lists: queues, stacks, activity feeds. Sets: unique members, tags, user followers (SADD, SMEMBERS, SINTER, SUNION). Sorted Sets: leaderboards, time-series, priority queues (ZADD, ZRANGE, ZREVRANK). Hashes: objects with fields, sessions, user profiles. Streams: event logs, message queues with consumer groups.