Cuckoo filter is a Bloom filter replacement for approximated set-membership queries. While Bloom filters are well-known space-efficient data structures to serve queries like "if item x is in a set?", they do not support deletion. Their variances to enable deletion (like counting Bloom filters) usually require much more space.
Cuckoo filters provide the flexibility to add and remove items dynamically. A cuckoo filter is based on cuckoo hashing (and therefore named as cuckoo filter). It is essentially a cuckoo hash table storing each key's fingerprint. Cuckoo hash tables can be highly compact, thus a cuckoo filter could use less space than conventional Bloom filters, for applications that require low false positive rates (< 3%).
For details about the algorithm and citations please use this article for now
"Cuckoo Filter: Better Than Bloom" by Bin Fan, Dave Andersen and Michael Kaminsky
The paper cited above leaves several parameters to choose. In this implementation
1 and 2 are suggested to be the optimum by the authors. The choice of 3 comes down to the desired false positive rate. Given a target false positive rate of r
and a bucket size b
, they suggest choosing the fingerprint size f
using
f >= log2(2b/r) bits
With the 8 bit fingerprint size in this repository, you can expect r ~= 0.03
.
Other implementations use 16 bit, which correspond to a false positive rate of r ~= 0.0001
.
package main
import "fmt"
import cuckoo "github.com/seiflotfy/cuckoofilter"
func main() {
cf := cuckoo.NewFilter(1000)
cf.InsertUnique([]byte("geeky ogre"))
// Lookup a string (and it a miss) if it exists in the cuckoofilter
cf.Lookup([]byte("hello"))
count := cf.Count()
fmt.Println(count) // count == 1
// Delete a string (and it a miss)
cf.Delete([]byte("hello"))
count = cf.Count()
fmt.Println(count) // count == 1
// Delete a string (a hit)
cf.Delete([]byte("geeky ogre"))
count = cf.Count()
fmt.Println(count) // count == 0
cf.Reset() // reset
}