Skip to content
This repository has been archived by the owner. It is now read-only.
Go to file

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

Update: this project has been suspended and superceded by

countminsketch: an approximate items counter Redis module

This is trivial implementation of the Count Min Sketch (CMS) from Graham Cormode and S. Muthukrishnan as described in (Approximating Data with the Count-Min Data Structure)[].

CMS is a probabilistic data structure that can be used to keep values associated with items, such as the counts of observations of samples in a stream (e.g. IP addresses, URLs, etc...). The counts are only approximated, i.e. there is a probability for error, but the data structure requires considerably less resources compared to storing all samples and their counts.

This implementation employs Redis' Strings for storing the data structure using direct memory access and 16-bit registers.

Quick start guide

  1. Build a Redis server with support for modules.
  2. Build the countminsketch module: make
  3. To load a module, Start Redis with the --loadmodule /path/to/ option, add it as a directive to the configuration file or send a MODULE LOAD command.


CMS.INCRBY key item value [item value ...]

Time complexity: O(N), where N is the number of items.

Increments the count for item by value.

If key does not exist, the sketch is initialized using a default width of 2000 and a depth of 10 (the defaults provide a %0.01 error at a probability of %0.01).

Reply: Simple String, "OK".

CMS.QUERY key item [item ...]

Time complexity: O(N), where N is the number of items.

Returns the estimated count for item.

Reply: Array of Integers, nil if key or element not found

CMS.MERGE destkey numkeys key [key ...] [WEIGHTS weight [weight ...]]

Time complexity: O(N), where N is the number of keys.

Merges numkeys of key into destkey. By default, destkey holds the sum of the sketches associated with the specified key, where all key should exist and have the same width and depth. Like Redis's builtin ZUNIONSTORE and ZINTERSTORE commands, it is possible to weight each sketch using the WEIGHTS option.

Reply: Simple String, "OK".

CMS.INITBYDIM key width depth

Time complexity: O(1)

Initializes a CMS to the dimensions specified by width and depth.

Reply: Simple String, "OK".

CMS.INITBYERR key error probabilty

Time complexity: O(1)

Initializes a CMS with a maximum error at a given probability.

error and probability should be given as doubles. For example, to specify an error of at most %0.01 at probability of %0.01, use CMS.INITBYERR key 0.001 0.001.

Reply: Simple String, "OK".


Time complexity: O(1)

Debugging helper - shows the sketch's properties.

Reply: Bulk Array.


Issue reports, pull and feature requests are welcome.



Uses Redis Copyright (C) 2009-2012, Salvatore Sanfilippo, BSD 3-Clause License. Uses xxHash Copyright (C) 2012-2014, Yann Collet, BSD 2-Clause License.


An apporximate frequency counter Redis module




No releases published


No packages published

Contributors 4

You can’t perform that action at this time.