Redis hyperloglog cardinality statistics algorithm

Use software :redis

edition ：2.8.9

effect ： Add all element parameters to the... Stored at the variable name specified as the first parameter HyperLogLog In the data structure

HLL It has the following characteristics ：

• The ability to use very little memory to count huge amounts of data , It just needs 12K Space can be counted 2^64 The data of ;
• There is a certain error in statistics , The error rate is low overall , The standard error is 0.81%;
• The error can be reduced by setting an auxiliary calculation factor .

Additive elements ：  This command supports adding one or more elements to HLL In structure .

127.0.0.1:6379> pfadd key "redis"
(integer) 1
127.0.0.1:6379> pfadd key "java" "sql"
(integer) 1

Count non repeating elements ：

127.0.0.1:6379> pfadd key "redis"
(integer) 1
127.0.0.1:6379> pfadd key "sql"
(integer) 1
127.0.0.1:6379> pfadd key "redis"
(integer) 0
127.0.0.1:6379> pfcount key
(integer) 2

Merge one or more HLL To new structure ：

127.0.0.1:6379> pfadd k "java" "sql"
(integer) 1
127.0.0.1:6379> pfadd k2 "redis" "sql"
(integer) 1
127.0.0.1:6379> pfmerge k3 k k2
OK
127.0.0.1:6379> pfcount k3
(integer) 3