A HyperLogLog in Redis is an algorithm that allows you to count the number of unique items in a set without incurring significant memory usage. It works closely similar to a Redis filter bloom but with a different implementation.
There are three main commands when working with Redis HyperLogLogs. These commands include:
- PFADD command
- PFCOUNT command
- PFMERGE command.
Let us take an example.
PFADD Command
Assume we have a database that holds the type of databases we support. We can add each entry to the HyperLogLog as:
127.0.0.1:6379> PFADD databases MySQL
(integer) 1
127.0.0.1:6379> PFADD databases MongoDB
(integer) 1
127.0.0.1:6379> PFADD databases PostgreSQL
(integer) 1
127.0.0.1:6379> PFADD databases Oracle
(integer) 1
(integer) 1
127.0.0.1:6379> PFADD databases MongoDB
(integer) 1
127.0.0.1:6379> PFADD databases PostgreSQL
(integer) 1
127.0.0.1:6379> PFADD databases Oracle
(integer) 1
PFCOUNT Command
To see the number of databases supported, we can run:
127.0.0.1:6379> PFCOUNT databases
(integer) 4
(integer) 4
PFMERGE Command
Let us also assume that we have a list of unsupported databases:
127.0.0.1:6379> PFADD unsupported Elasticsearch Solr Neo4j Memcached
(integer) 1
(integer) 1
To create a union of both supported and unsupported databases, we can use the PFMERGE command as shown:
127.0.0.1:6379> PFMERGE all databases unsupported
OK
OK
We can then count the number of databases in the new key is:
127.0.0.1:6379> PFCOUNT all
(integer) 8
(integer) 8
Closing
This tutorial covered how to work with Redis HyperLogLog using PFADD, PFCOUNT, and PFMERGE commands.
