README
¶
Quotient Filter
a go implementation of the quotient filter data structure. Quotient filter is a probabilistic data structure similar to a Bloom Filter. As BF it is used to approximate set membership. It never returns false for added keys but has a probability of returning a false positive, the probability depends on the amount of remainder bits and fill rate of the filter.
Its size is 10-20% more than a bloom filter with same FP rate but it usually is faster. It also has a few additional benefits:
- Filters can be merged without rehashing keys
- Adding or checking a key requires evaluating only a single hash function
- Can be resized without rehashing
For more details, see: https://en.wikipedia.org/wiki/Quotient_filter
Interface
// Create a filter that can hold 1m elements while maintaining 1% false positive
// rate when at 1 million items length.
qf := NewProbability(1000000, 0.01)
qf.Add("key")
if !qf.Contains("key") {
panic("False negative not possible")
}
docs
https://godoc.org/github.com/Nomon/qf-go
Credits
All credit goes to https://github.com/vedantk/quotient-filter this repo is golang port of Vedant Kumars C implementatuon.
License MIT
Additional work Copyright (c) 2016 Matti Savolainen matti@unity3d.com
Copyright 2014 Vedant Kumar vsk@berkeley.edu
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE..
Documentation
¶
Index ¶
Constants ¶
This section is empty.
Variables ¶
var ErrFull = errors.New("filter is at its max capacity")
ErrFull is returned when Add is called while the filter is at max capacity.
Functions ¶
This section is empty.
Types ¶
type QuotientFilter ¶
type QuotientFilter struct {
// contains filtered or unexported fields
}
QuotientFilter is a basic quotient filter implementation. None of the methods are thread safe.
func New ¶
func New(q, r uint8) *QuotientFilter
New returns a QuotientFilter with q quotient bits and r remainder bits. it can hold 1 << q elements.
func NewHash ¶
func NewHash(h hash.Hash64, q, r uint8) *QuotientFilter
NewHash returns a QuotientFilter backed by a different hash function. Default hash function is FNV-64a
func NewProbability ¶
func NewProbability(capacity int, probability float64) *QuotientFilter
NewProbability returns a quotient filter that can accomidate capacity number of elements and maintain the probability passed.
func (*QuotientFilter) Add ¶
func (qf *QuotientFilter) Add(key string) error
Add adds the key to the filter.
func (*QuotientFilter) AddAll ¶
func (qf *QuotientFilter) AddAll(keys []string) error
AddAll adds multiple keys to the filter
func (*QuotientFilter) Contains ¶
func (qf *QuotientFilter) Contains(key string) bool
Contains checks if key is present in the filter false positive probability is based on q, r and number of added keys false negatives are not possible, unless Delete is used in conjunction with a hash function that yields more that q+r bits.
func (*QuotientFilter) FPProbability ¶
func (qf *QuotientFilter) FPProbability() float64
FPProbability returns the probability for false positive with the current fillrate n = length m = capacity a = n / m r = remainder bits then probability for false positive is 1 - e^(-a/2^r) <= 2^-r
Source Files