Skip to content

data-structures.md

Latest commit

 

History

History
39 lines (19 loc) · 1.51 KB

data-structures.md

File metadata and controls

39 lines (19 loc) · 1.51 KB

Redis Data Structures

At root, its a key-val in-mem db written in C. Also allows memory snapshots as AOF (or RDB) for persistence.

Basic Value Types

  • Primitve string type. If required data structure ain't supported by Redis, can always just serialize and use this. Max length 512MB.

  • Hashes for shallow objects.

  • Lists with items sorted by insertion order. Implemented with LinkedLists, so good for Queues but positional index is slower (better suited to Sorted Sets).

  • Sets, only store unique elements & do allow union/intersection/other-op.

  • Sorted Sets, ordered on (explicit) score & lexical order. O(N) for add/del.

  • GeoSpatial Index Support based on Sorted Set, with score based on Geohash algorithm.

Probabilistic

  • Hyperloglogs to get quick estimate of unique value in a huge dataset.

Streams

  • This log data-structure is append-only table, allowing multiple consumers to track event data.

More About Strings

  • Bit-oriented op are supported on Redis strings. So, for 512MB it can manipulate 2^32 bits. Eg: usage like tracking up/down states of million sensors in a simple bitmapped string.

  • Could also be used to cache requests/configs/states, have counters. Can have expiry on key.

  • String Ops with O(N) are APPEND, BITCOUNT, BITOP, BITPOS, GETRANGE, MGET, MSET, MSERTNX.

  • String Ops with O(1) are BITFIELD, DECR, DECRBY, GET, GETBIT, GETSET, INCR, INCRBY, INCRBYFLOAT, PSETEX, SET, SETBIT, SETEX, SETNX, SETRANGE, SETRLEN.