6 Commits

Author SHA1 Message Date
46b59d743e feat: custom lint rule, fixed readme 2026-04-29 20:54:05 -04:00
39548b4332 feat!: Drop returns bool, Put doesn't stack-overflow (#21)
All checks were successful
CI / Check PR Title (push) Has been skipped
CI / Go Lint (push) Successful in 58s
CI / Makefile Lint (push) Successful in 55s
CI / Markdown Lint (push) Successful in 34s
CI / Unit Tests (push) Successful in 54s
CI / Fuzz Tests (push) Successful in 1m26s
CI / Mutation Tests (push) Successful in 1m11s
## Description

Closes #11.

## Changes

### Design Decisions

## Checklist

- [ ] Tests pass
- [ ] Docs updated

Reviewed-on: #21
Co-authored-by: M.V. Hutz <git@maximhutz.me>
Co-committed-by: M.V. Hutz <git@maximhutz.me>
2026-04-17 01:31:01 +00:00
29ba6bfd4d fix!: no mixed receiver types (#23)
All checks were successful
CI / Check PR Title (push) Has been skipped
CI / Makefile Lint (push) Successful in 50s
CI / Go Lint (push) Successful in 54s
CI / Markdown Lint (push) Successful in 50s
CI / Unit Tests (push) Successful in 47s
CI / Mutation Tests (push) Successful in 1m31s
CI / Fuzz Tests (push) Successful in 1m21s
## Description

Currently, `bucket` and `Table` have mixed receiver types: some are pointer receviers, and others are value receivers.

As per the Go Wiki, [you can have value and pointer receivers, just don't mix them](https://go.dev/doc/faq#methods_on_values_or_pointers).

## Changes

- Replace all value receivers in `bucket` and `Table` with pointer receivers.

### Design Decisions

This decision was made due to the advice on the Go wiki.

## Checklist

- [x] Tests pass
- [x] Docs updated

Reviewed-on: #23
Co-authored-by: M.V. Hutz <git@maximhutz.me>
Co-committed-by: M.V. Hutz <git@maximhutz.me>
2026-04-16 03:27:48 +00:00
7cc1657403 refactor!: shorter constructors, bucketsubtable (#22)
All checks were successful
CI / Check PR Title (push) Has been skipped
CI / Makefile Lint (push) Successful in 47s
CI / Go Lint (push) Successful in 51s
CI / Markdown Lint (push) Successful in 46s
CI / Unit Tests (push) Successful in 47s
CI / Fuzz Tests (push) Successful in 1m19s
CI / Mutation Tests (push) Successful in 1m36s
## Description

Currently, the name of `bucket` is a bit confusing, because it is considered a 'table' in literature (as well as the whole hash table). A `bucket` is better described as a 'subtable', which is used by the total hash table to perform cuckoo hashing.

In addition, the constructors `NewTable`, `NewTableBy`, and `NewCustomTable` were given shorter names, because the package name `cuckoo` already implies that `New*` would create a hash table with cuckoo hashing. This package has one use-case, and so it unambiguous what constructors produce.

## Changes

- `NewTable` -> `New`
- `NewTableBy` -> `NewBy`
- `NewCustomTable` -> `NewCustom`
- `bucket` -> `subtable`

### Design Decisions

- I would have renamed `Table` and `subtable` to map equivalents, but 'submap' implies that a certain subsection of the map is contained within it, which isn't quite right.
- I chose not to go with `Map` and `table`, because of the split naming convention.

## Checklist

- [x] Tests pass
- [x] Docs updated

Reviewed-on: #22
Co-authored-by: M.V. Hutz <git@maximhutz.me>
Co-committed-by: M.V. Hutz <git@maximhutz.me>
2026-04-16 03:15:39 +00:00
42c5b5f8f4 feat!: update get from (V, error) to (V, bool) (#20)
All checks were successful
CI / Check PR Title (push) Has been skipped
CI / Go Lint (push) Successful in 43s
CI / Makefile Lint (push) Successful in 41s
CI / Markdown Lint (push) Successful in 32s
CI / Unit Tests (push) Successful in 39s
CI / Fuzz Tests (push) Successful in 1m44s
CI / Mutation Tests (push) Successful in 1m28s
## Description

Currently, the signature for `Table.Get` is `func (K) (V, error)`. This is not very Go-idiomatic, which prefers to return a boolean instead of an error. For instance, a built-in Go map is used like so:

```go
if value, ok := users[id]; !ok {
  // ...
}
```

Updating our table to look like that is best practice. In that same vein, to support direct lookup (i.e. `v := users[id]`), this PR also adds `Table.Find`.

## Changes

- BREAKING CHANGE: Update contract of `Table.Get` to `func (K) (V, bool)`. Returns 'false' is the item cannot be found, and 'true' if it is found.
- Add `Table.Find`.
- Updated tests and documentation to match the change.

### Design Decisions

- Chose to make this decision because throwing an error implies that there is something 'wrong' with the table. There is nothing wrong with the table; it is just that the item does not exist.

## Checklist

- [x] Tests pass
- [x] Docs updated

Reviewed-on: #20
Co-authored-by: M.V. Hutz <git@maximhutz.me>
Co-committed-by: M.V. Hutz <git@maximhutz.me>
2026-04-14 01:58:15 +00:00
867a1d49df feat: sentinel error ErrBadHash (#19)
All checks were successful
CI / Check PR Title (push) Has been skipped
CI / Makefile Lint (push) Successful in 1m4s
CI / Markdown Lint (push) Successful in 32s
CI / Go Lint (push) Successful in 1m15s
CI / Unit Tests (push) Successful in 38s
CI / Fuzz Tests (push) Successful in 1m34s
CI / Mutation Tests (push) Successful in 1m31s
## Description

Currently, the errors are not sentinel, and so are hard to test for. This PR makes sure hash collision errors are accounted for.

## Changes

- Add `ErrBadHash`. Happens when there are too many collisions for an item to be added.

### Design Decisions

- Chose to name `ErrBadHash` over `ErrCycle` because the feedbach that the user should be given is to evaluate their hash functions. Cycle collision is a bit esoteric.

## Checklist

- [x] Tests pass
- [x] Docs updated

Reviewed-on: #19
Co-authored-by: M.V. Hutz <git@maximhutz.me>
Co-committed-by: M.V. Hutz <git@maximhutz.me>
2026-04-14 00:38:11 +00:00
16 changed files with 315 additions and 251 deletions

View File

@@ -114,6 +114,9 @@ linters:
# Reports uses of functions with replacement inside the testing package.
- usetesting
# Reports mixed receiver types in structs/interfaces.
- recvcheck
settings:
revive:
rules:

14
.markdownlint-cli2.jsonc Normal file
View File

@@ -0,0 +1,14 @@
{
"customRules": [".markdownlint-rules/one-sentence-per-line.mjs"],
"config": {
"default": true,
"heading-style": { "style": "atx" },
"ul-indent": { "indent": 2 },
"line-length": false,
"no-duplicate-heading": { "siblings_only": true },
"no-inline-html": {
"allowed_elements": ["br", "code", "details", "summary", "img", "picture", "source"]
},
"first-line-heading": true
}
}

View File

@@ -0,0 +1,48 @@
// @ts-check
/**
* @typedef {object} ErrorInfo
* @property {number} lineNumber
* @property {string} [context]
* @property {string} [detail]
* @property {[number, number]} [range]
*/
/**
* @typedef {object} Params
* @property {string[]} lines
*/
/** @typedef {(error: ErrorInfo) => void} OnError */
/**
* @typedef {object} Rule
* @property {string[]} names
* @property {string} description
* @property {string[]} tags
* @property {(params: Params, onError: OnError) => void} function
*/
/** @type {Rule} */
export default {
names: ["one-sentence-per-line"],
description: "Each sentence must be on its own line",
tags: ["sentences"],
function: function (params, onError) {
let inFence = false;
params.lines.forEach((line, index) => {
if (/^```/.test(line)) {
inFence = !inFence;
return;
}
if (inFence) return;
// Skip headings, blank lines, HTML, table rows
if (/^(#|\s*[|<]|>|\s*$)/.test(line)) return;
// Strip list marker before checking
const text = line.replace(/^\s*(?:[-*+]|\d+\.)\s+/, "");
if (/[.!?]\s+[A-Z]/.test(text)) {
onError({ lineNumber: index + 1, context: text.trim() });
}
});
},
};

View File

@@ -1,17 +0,0 @@
default: true
heading-style:
style: atx
ul-indent:
indent: 2
line-length: false
no-duplicate-heading:
siblings_only: true
no-inline-html:
allowed_elements:
- br
- details
- summary
- img
- picture
- source
first-line-heading: true

View File

@@ -1,3 +1,6 @@
# <img height="30" src="assets/logo.svg" alt="Go Cuckoo, by `mvhutz`."> Go Cuckoo
A hash table that uses cuckoo hashing to achieve a worst-case O(1) lookup time. Read more about it in [the package documentation](https://pkg.go.dev/git.maximhutz.com/tools/go-cuckoo).
A hash table that uses cuckoo hashing to achieve a worst-case O(1) lookup time.
Read more about it in [the package documentation][docs].
[docs]: https://pkg.go.dev/git.maximhutz.com/tools/go-cuckoo

107
bucket.go
View File

@@ -1,107 +0,0 @@
package cuckoo
// An Entry is a key-value pair.
type Entry[K, V any] struct {
Key K
Value V
}
type slot[K, V any] struct {
Entry[K, V]
occupied bool
}
type bucket[K, V any] struct {
hash Hash[K]
slots []slot[K, V]
capacity, size uint64
compare EqualFunc[K]
}
// location determines where in the bucket a certain key would be placed. If the
// capacity is 0, this will panic.
func (b bucket[K, V]) location(key K) uint64 {
return b.hash(key) % b.capacity
}
func (b bucket[K, V]) get(key K) (value V, found bool) {
if b.capacity == 0 {
return
}
slot := b.slots[b.location(key)]
return slot.Value, slot.occupied && b.compare(slot.Key, key)
}
func (b *bucket[K, V]) drop(key K) (occupied bool) {
if b.capacity == 0 {
return
}
slot := &b.slots[b.location(key)]
if slot.occupied && b.compare(slot.Key, key) {
slot.occupied = false
b.size--
return true
}
return false
}
func (b *bucket[K, V]) resized(capacity uint64) bucket[K, V] {
return bucket[K, V]{
slots: make([]slot[K, V], capacity),
capacity: capacity,
hash: b.hash,
compare: b.compare,
}
}
func (b bucket[K, V]) update(key K, value V) (updated bool) {
if b.capacity == 0 {
return
}
slot := &b.slots[b.location(key)]
if slot.occupied && b.compare(slot.Key, key) {
slot.Value = value
return true
}
return false
}
func (b *bucket[K, V]) insert(insertion Entry[K, V]) (evicted Entry[K, V], eviction bool) {
if b.capacity == 0 {
return insertion, true
}
slot := &b.slots[b.location(insertion.Key)]
if !slot.occupied {
slot.Entry = insertion
slot.occupied = true
b.size++
return
}
if b.compare(slot.Key, insertion.Key) {
slot.Value = insertion.Value
return
}
insertion, slot.Entry = slot.Entry, insertion
return insertion, true
}
func newBucket[K, V any](capacity uint64, hash Hash[K], compare EqualFunc[K]) bucket[K, V] {
return bucket[K, V]{
hash: hash,
capacity: capacity,
compare: compare,
size: 0,
slots: make([]slot[K, V], capacity),
}
}

View File

@@ -2,7 +2,7 @@ package cuckoo
// An EqualFunc determines whethers two keys are 'equal'. Keys that are 'equal'
// are teated as the same by the [Table]. A good EqualFunc is pure,
// deterministic, and fast. By default, [NewTable] uses [DefaultEqualFunc].
// deterministic, and fast. By default, [New] uses [DefaultEqualFunc].
//
// This function MUST NOT return true if the [Hash] digest of two keys
// are different: the [Table] will not work.

View File

@@ -28,7 +28,7 @@ func ExampleEqualFunc_badEqualFunc() {
// Two users with the same ID are equal.
isEqual := func(a, b User) bool { return a.ID == b.ID }
userbase := cuckoo.NewCustomTable[User, bool](makeHash(1), makeHash(2), isEqual)
userbase := cuckoo.NewCustom[User, bool](makeHash(1), makeHash(2), isEqual)
(userbase.Put(User{"1", "Robert Doe"}, true))

View File

@@ -56,7 +56,7 @@ func FuzzInsertLookup(f *testing.F) {
fmt.Fprintf(os.Stderr, "seedA=%d seedB=%d capacity=%d growthFactor=%d\n",
seedA, seedB, capacity, growthFactor)
actual := cuckoo.NewCustomTable[uint32, uint32](
actual := cuckoo.NewCustom[uint32, uint32](
offsetHash(seedA),
offsetHash(seedB),
func(a, b uint32) bool { return a == b },
@@ -74,16 +74,16 @@ func FuzzInsertLookup(f *testing.F) {
delete(expected, step.key)
_, err := actual.Get(step.key)
assert.Error(err)
_, ok = actual.Get(step.key)
assert.False(ok)
} else {
_, err := actual.Put(step.key, step.value)
err := actual.Put(step.key, step.value)
assert.NoError(err)
expected[step.key] = step.value
found, err := actual.Get(step.key)
assert.NoError(err)
found, ok := actual.Get(step.key)
assert.True(ok)
assert.Equal(step.value, found)
}

View File

@@ -11,7 +11,7 @@ func TestMaxEvictions(t *testing.T) {
assert := assert.New(t)
for i := 16; i < 116; i++ {
table := NewTable[int, bool](Capacity(i / 2))
table := New[int, bool](Capacity(i / 2))
expectedEvictions := 3 * math.Floor(math.Log2(float64(i)))
assert.Equal(table.maxEvictions(), int(expectedEvictions))
@@ -20,10 +20,10 @@ func TestMaxEvictions(t *testing.T) {
func TestLoad(t *testing.T) {
assert := assert.New(t)
table := NewTable[int, bool](Capacity(8))
table := New[int, bool](Capacity(8))
for i := range 16 {
_, err := table.Put(i, true)
err := table.Put(i, true)
assert.NoError(err)
assert.Equal(float64(table.Size())/float64(table.TotalCapacity()), table.load())
}

View File

@@ -14,7 +14,7 @@ import (
func TestNewTable(t *testing.T) {
assert := assert.New(t)
table := cuckoo.NewTable[int, bool]()
table := cuckoo.New[int, bool]()
assert.NotNil(table)
assert.Zero(table.Size())
@@ -23,9 +23,9 @@ func TestNewTable(t *testing.T) {
func TestAddItem(t *testing.T) {
assert := assert.New(t)
key, value := 0, true
table := cuckoo.NewTable[int, bool]()
table := cuckoo.New[int, bool]()
_, err := table.Put(key, value)
err := table.Put(key, value)
assert.NoError(err)
assert.Equal(1, table.Size())
@@ -35,10 +35,10 @@ func TestAddItem(t *testing.T) {
func TestPutOverwrite(t *testing.T) {
assert := assert.New(t)
key, value, newValue := 0, 1, 2
table := cuckoo.NewTable[int, int]()
table := cuckoo.New[int, int]()
(table.Put(key, value))
_, err := table.Put(key, newValue)
err := table.Put(key, newValue)
assert.NoError(err)
assert.Equal(1, table.Size())
@@ -50,11 +50,11 @@ func TestPutOverwrite(t *testing.T) {
func TestSameHash(t *testing.T) {
assert := assert.New(t)
hash := func(int) uint64 { return 0 }
table := cuckoo.NewCustomTable[int, bool](hash, hash, cuckoo.DefaultEqualFunc[int])
table := cuckoo.NewCustom[int, bool](hash, hash, cuckoo.DefaultEqualFunc[int])
_, errA := table.Put(0, true)
_, errB := table.Put(1, true)
_, errC := table.Put(2, true)
errA := table.Put(0, true)
errB := table.Put(1, true)
errC := table.Put(2, true)
assert.NoError(errA)
assert.NoError(errB)
@@ -63,20 +63,20 @@ func TestSameHash(t *testing.T) {
func TestStartingCapacity(t *testing.T) {
assert := assert.New(t)
table := cuckoo.NewTable[int, bool](cuckoo.Capacity(64))
table := cuckoo.New[int, bool](cuckoo.Capacity(64))
assert.Equal(uint64(128), table.TotalCapacity())
}
func TestResizeCapacity(t *testing.T) {
assert := assert.New(t)
table := cuckoo.NewTable[int, bool](
table := cuckoo.New[int, bool](
cuckoo.Capacity(8),
cuckoo.GrowthFactor(2),
)
for table.TotalCapacity() == 16 {
_, err := table.Put(rand.Int(), true)
err := table.Put(rand.Int(), true)
assert.NoError(err)
}
@@ -85,11 +85,11 @@ func TestResizeCapacity(t *testing.T) {
func TestPutMany(t *testing.T) {
assert := assert.New(t)
expected, actual := map[int]bool{}, cuckoo.NewTable[int, bool]()
expected, actual := map[int]bool{}, cuckoo.New[int, bool]()
for i := range 1_000 {
expected[i] = true
_, err := actual.Put(i, true)
err := actual.Put(i, true)
assert.NoError(err)
}
@@ -100,20 +100,20 @@ func TestPutMany(t *testing.T) {
func TestGetMany(t *testing.T) {
assert := assert.New(t)
table := cuckoo.NewTable[int, bool]()
table := cuckoo.New[int, bool]()
for i := range 1_000 {
_, err := table.Put(i, true)
err := table.Put(i, true)
assert.NoError(err)
}
for i := range 2_000 {
value, err := table.Get(i)
value, ok := table.Get(i)
if i < 1_000 {
assert.NoError(err)
assert.True(ok)
assert.Equal(value, true)
} else {
assert.Error(err)
assert.False(ok)
}
}
}
@@ -121,7 +121,7 @@ func TestGetMany(t *testing.T) {
func TestDropExistingItem(t *testing.T) {
assert := assert.New(t)
key, value := 0, true
table := cuckoo.NewTable[int, bool]()
table := cuckoo.New[int, bool]()
(table.Put(key, value))
had := table.Drop(key)
@@ -134,7 +134,7 @@ func TestDropExistingItem(t *testing.T) {
func TestDropNoItem(t *testing.T) {
assert := assert.New(t)
key := 0
table := cuckoo.NewTable[int, bool]()
table := cuckoo.New[int, bool]()
had := table.Drop(key)
@@ -146,7 +146,7 @@ func TestDropNoItem(t *testing.T) {
func TestDropItemCapacity(t *testing.T) {
assert := assert.New(t)
key := 0
table := cuckoo.NewTable[int, bool](
table := cuckoo.New[int, bool](
cuckoo.Capacity(64),
cuckoo.GrowthFactor(2),
)
@@ -163,11 +163,11 @@ func TestDropItemCapacity(t *testing.T) {
func TestPutNoCapacity(t *testing.T) {
assert := assert.New(t)
key, value := 0, true
table := cuckoo.NewTable[int, bool](
table := cuckoo.New[int, bool](
cuckoo.Capacity(0),
)
_, err := table.Put(key, value)
err := table.Put(key, value)
assert.NoError(err)
assert.Equal(1, table.Size())
@@ -176,16 +176,16 @@ func TestPutNoCapacity(t *testing.T) {
func TestBadHashCapacity(t *testing.T) {
assert := assert.New(t)
table := cuckoo.NewCustomTable[int, bool](
table := cuckoo.NewCustom[int, bool](
func(int) uint64 { return 0 },
func(int) uint64 { return 0 },
func(a, b int) bool { return a == b },
cuckoo.Capacity(20),
)
_, err1 := table.Put(0, true)
_, err2 := table.Put(1, true)
_, err3 := table.Put(2, true)
err1 := table.Put(0, true)
err2 := table.Put(1, true)
err3 := table.Put(2, true)
assert.NoError(err1)
assert.NoError(err2)
@@ -196,12 +196,12 @@ func TestBadHashCapacity(t *testing.T) {
func TestDropResizeCapacity(t *testing.T) {
assert := assert.New(t)
table := cuckoo.NewTable[int, bool](
table := cuckoo.New[int, bool](
cuckoo.Capacity(10),
)
_, err1 := table.Put(0, true)
_, err2 := table.Put(1, true)
err1 := table.Put(0, true)
err2 := table.Put(1, true)
table.Drop(1)
assert.NoError(errors.Join(err1, err2))
@@ -216,11 +216,9 @@ func TestNewTableBy(t *testing.T) {
}
assert := assert.New(t)
table := cuckoo.NewTableBy[User, bool](
func(u User) string { return u.id },
)
table := cuckoo.NewBy[User, bool](func(u User) string { return u.id })
_, err := table.Put(User{nil, "1", "Robert"}, true)
err := table.Put(User{nil, "1", "Robert"}, true)
assert.NoError(err)
assert.Equal(1, table.Size())

7
doc.go
View File

@@ -1,9 +1,12 @@
// Package cuckoo provides a hash table that uses cuckoo hashing to achieve
// a worst-case O(1) lookup time.
//
// While a [NewTable] only supports comparable keys by default, you can create
// a table with any key type using [NewCustomTable]. Custom [Hash] functions and
// While a [New] only supports comparable keys by default, you can create
// a table with any key type using [NewCustom]. Custom [Hash] functions and
// key comparison are also supported.
//
// NOTE: The [Table] is a look-up structure, and not a source of truth. If
// [ErrBadHash] occurs, the data cannot be restored.
//
// See more: https://en.wikipedia.org/wiki/Cuckoo_hashing
package cuckoo

View File

@@ -8,25 +8,25 @@ import (
)
func Example_basic() {
table := cuckoo.NewTable[int, string]()
table := cuckoo.New[int, string]()
if _, err := table.Put(1, "Hello, World!"); err != nil {
if err := table.Put(1, "Hello, World!"); err != nil {
fmt.Println("Put error:", err)
}
if item, err := table.Get(1); err != nil {
fmt.Println("Error:", err)
if item, ok := table.Get(1); !ok {
fmt.Println("Not Found 1!")
} else {
fmt.Println("Found 1:", item)
}
if item, err := table.Get(0); err != nil {
fmt.Println("Error:", err)
if item, ok := table.Get(0); !ok {
fmt.Println("Not Found 0!")
} else {
fmt.Println("Found 0:", item)
}
// Output:
// Found 1: Hello, World!
// Error: key '0' not found
// Not Found 0!
}

View File

@@ -9,7 +9,7 @@ import "fmt"
const DefaultCapacity uint64 = 16
// DefaultGrowthFactor is the standard resize multiplier for a [Table]. Most
// hash table implementations use 2.
// implementations use 2.
const DefaultGrowthFactor uint64 = 2
// defaultMinimumLoad is the default lowest acceptable occupancy of a [Table].
@@ -31,10 +31,10 @@ type settings struct {
}
// An Option modifies the settings of a [Table]. It is used in its constructors
// like [NewTable], for example.
// like [New], for example.
type Option func(*settings)
// Capacity modifies the starting capacity of each bucket of the [Table]. The
// Capacity modifies the starting capacity of each subtable of the [Table]. The
// value must be non-negative.
func Capacity(value int) Option {
if value < 0 {

107
subtable.go Normal file
View File

@@ -0,0 +1,107 @@
package cuckoo
// An entry is a key-value pair.
type entry[K, V any] struct {
key K
value V
}
type slot[K, V any] struct {
entry[K, V]
occupied bool
}
type subtable[K, V any] struct {
hash Hash[K]
slots []slot[K, V]
capacity, size uint64
compare EqualFunc[K]
}
// location determines where in the subtable a certain key would be placed. If
// the capacity is 0, this will panic.
func (t *subtable[K, V]) location(key K) uint64 {
return t.hash(key) % t.capacity
}
func (t *subtable[K, V]) get(key K) (value V, found bool) {
if t.capacity == 0 {
return
}
slot := t.slots[t.location(key)]
return slot.value, slot.occupied && t.compare(slot.key, key)
}
func (t *subtable[K, V]) drop(key K) (occupied bool) {
if t.capacity == 0 {
return
}
slot := &t.slots[t.location(key)]
if slot.occupied && t.compare(slot.key, key) {
slot.occupied = false
t.size--
return true
}
return false
}
func (t *subtable[K, V]) resized(capacity uint64) *subtable[K, V] {
return &subtable[K, V]{
slots: make([]slot[K, V], capacity),
capacity: capacity,
hash: t.hash,
compare: t.compare,
}
}
func (t *subtable[K, V]) update(key K, value V) (updated bool) {
if t.capacity == 0 {
return
}
slot := &t.slots[t.location(key)]
if slot.occupied && t.compare(slot.key, key) {
slot.value = value
return true
}
return false
}
func (t *subtable[K, V]) insert(insertion entry[K, V]) (evicted entry[K, V], eviction bool) {
if t.capacity == 0 {
return insertion, true
}
slot := &t.slots[t.location(insertion.key)]
if !slot.occupied {
slot.entry = insertion
slot.occupied = true
t.size++
return
}
if t.compare(slot.key, insertion.key) {
slot.value = insertion.value
return
}
insertion, slot.entry = slot.entry, insertion
return insertion, true
}
func newSubtable[K, V any](capacity uint64, hash Hash[K], compare EqualFunc[K]) *subtable[K, V] {
return &subtable[K, V]{
hash: hash,
capacity: capacity,
compare: compare,
size: 0,
slots: make([]slot[K, V], capacity),
}
}

146
table.go
View File

@@ -1,41 +1,50 @@
package cuckoo
import (
"errors"
"fmt"
"iter"
"math/bits"
"strings"
)
// A Table is hash table that uses cuckoo hashing to resolve collision. Create
// one with [NewTable]. Or if you want more granularity, use [NewTableBy] or
// [NewCustomTable].
// ErrBadHash occurs when the hashes given to a [Table] cause too many key
// collisions. Discard the old table, rebuild it from your source data, and try:
//
// 1. Different hash seeds. Equal seeds produce equal hash functions, which
// always cycle.
// 2. A different [Hash] algorithm.
var ErrBadHash = errors.New("bad hash")
// A Table which uses cuckoo hashing to resolve collision. Create
// one with [New]. Or if you want more granularity, use [NewBy] or
// [NewCustom].
type Table[K, V any] struct {
bucketA, bucketB bucket[K, V]
growthFactor uint64
minLoadFactor float64
tableA, tableB *subtable[K, V]
growthFactor uint64
minLoadFactor float64
}
// TotalCapacity returns the number of slots allocated for the [Table]. To get the
// number of slots filled, look at [Table.Size].
func (t Table[K, V]) TotalCapacity() uint64 {
return t.bucketA.capacity + t.bucketB.capacity
func (t *Table[K, V]) TotalCapacity() uint64 {
return t.tableA.capacity + t.tableB.capacity
}
// Size returns how many slots are filled in the [Table].
func (t Table[K, V]) Size() int {
return int(t.bucketA.size + t.bucketB.size)
func (t *Table[K, V]) Size() int {
return int(t.tableA.size + t.tableB.size)
}
func log2(n uint64) (m int) {
return max(0, bits.Len64(n)-1)
}
func (t Table[K, V]) maxEvictions() int {
func (t *Table[K, V]) maxEvictions() int {
return 3 * log2(t.TotalCapacity())
}
func (t Table[K, V]) load() float64 {
func (t *Table[K, V]) load() float64 {
// When there are no slots in the table, we still treat the load as 100%.
// Every slot in the table is full.
if t.TotalCapacity() == 0 {
@@ -48,21 +57,21 @@ func (t Table[K, V]) load() float64 {
// insert attempts to put/update an entry in the table, without modifying the
// size of the table. Returns a displaced entry and 'homeless = true' if an
// entry could not be placed after exhausting evictions.
func (t *Table[K, V]) insert(entry Entry[K, V]) (displaced Entry[K, V], homeless bool) {
if t.bucketA.update(entry.Key, entry.Value) {
func (t *Table[K, V]) insert(entry entry[K, V]) (displaced entry[K, V], homeless bool) {
if t.tableA.update(entry.key, entry.value) {
return
}
if t.bucketB.update(entry.Key, entry.Value) {
if t.tableB.update(entry.key, entry.value) {
return
}
for range t.maxEvictions() {
if entry, homeless = t.bucketA.insert(entry); !homeless {
if entry, homeless = t.tableA.insert(entry); !homeless {
return
}
if entry, homeless = t.bucketB.insert(entry); !homeless {
if entry, homeless = t.tableB.insert(entry); !homeless {
return
}
}
@@ -72,12 +81,12 @@ func (t *Table[K, V]) insert(entry Entry[K, V]) (displaced Entry[K, V], homeless
// resized creates an empty copy of the table, with a new capacity for each
// bucket.
func (t Table[K, V]) resized(capacity uint64) Table[K, V] {
return Table[K, V]{
func (t *Table[K, V]) resized(capacity uint64) *Table[K, V] {
return &Table[K, V]{
growthFactor: t.growthFactor,
minLoadFactor: t.minLoadFactor,
bucketA: t.bucketA.resized(capacity),
bucketB: t.bucketB.resized(capacity),
tableA: t.tableA.resized(capacity),
tableB: t.tableB.resized(capacity),
}
}
@@ -88,16 +97,16 @@ func (t *Table[K, V]) resize(capacity uint64) bool {
updated := t.resized(capacity)
for k, v := range t.Entries() {
if _, failed := updated.insert(Entry[K, V]{k, v}); failed {
if _, failed := updated.insert(entry[K, V]{k, v}); failed {
return false
}
}
*t = updated
*t = *updated
return true
}
// grow increases the table's capacity by the [Table.growthFactor]. If the
// grow increases the table's capacity by the growth factor. If the
// capacity is 0, it increases it to 1.
func (t *Table[K, V]) grow() bool {
var newCapacity uint64
@@ -105,46 +114,49 @@ func (t *Table[K, V]) grow() bool {
if t.TotalCapacity() == 0 {
newCapacity = 1
} else {
newCapacity = t.bucketA.capacity * t.growthFactor
newCapacity = t.tableA.capacity * t.growthFactor
}
return t.resize(newCapacity)
}
// shrink reduces the table's capacity by the [Table.growthFactor]. It may
// shrink reduces the table's capacity by the growth factor. It may
// reduce it down to 0.
func (t *Table[K, V]) shrink() bool {
return t.resize(t.bucketA.capacity / t.growthFactor)
return t.resize(t.tableA.capacity / t.growthFactor)
}
// Get fetches the value for a key in the [Table]. Returns an error if no value
// is found.
func (t Table[K, V]) Get(key K) (value V, err error) {
if item, ok := t.bucketA.get(key); ok {
return item, nil
// Get fetches the value for a key in the [Table]. Matches the comma-ok pattern
// of a builtin map; see [Table.Find] for plain indexing.
func (t *Table[K, V]) Get(key K) (value V, ok bool) {
if item, ok := t.tableA.get(key); ok {
return item, true
}
if item, ok := t.bucketB.get(key); ok {
return item, nil
if item, ok := t.tableB.get(key); ok {
return item, true
}
return value, fmt.Errorf("key '%v' not found", key)
return
}
// Find fetches the value of a key. Matches direct indexing of a builtin map;
// see [Table.Get] for a comma-ok pattern.
func (t *Table[K, V]) Find(key K) (value V) {
value, _ = t.Get(key)
return
}
// Has returns true if a key has a value in the table.
func (t Table[K, V]) Has(key K) (exists bool) {
_, err := t.Get(key)
return err == nil
func (t *Table[K, V]) Has(key K) (exists bool) {
_, exists = t.Get(key)
return
}
// Put sets the value for a key. If it cannot be set, an error is returned,
// along with the last displaced entry.
//
// On failure, the returned entry and the current table contents together
// preserve all previously inserted entries and the attempted entry.
func (t *Table[K, V]) Put(key K, value V) (displaced Entry[K, V], err error) {
// Put sets the value for a key. If it cannot be set, an error is returned.
func (t *Table[K, V]) Put(key K, value V) (err error) {
var (
entry = Entry[K, V]{key, value}
entry = entry[K, V]{key, value}
homeless bool
)
@@ -157,24 +169,24 @@ func (t *Table[K, V]) Put(key K, value V) (displaced Entry[K, V], err error) {
// early when the table is sparse, while the latter catches cases where
// growing never helps.
if t.load() < t.minLoadFactor {
return entry, fmt.Errorf("bad hash: resize on load %d/%d", t.Size(), t.TotalCapacity())
return fmt.Errorf("hash functions produced a cycle at load %d/%d: %w", t.Size(), t.TotalCapacity(), ErrBadHash)
}
// It is theoretically possible to have a table with a larger capacity
// that is valid. But this chance is astronomically small, so we ignore
// it in this implementation.
if grew := t.grow(); !grew {
return entry, fmt.Errorf("bad hash: could not redistribute entries into larger table")
return fmt.Errorf("could not redistribute entries into larger table: %w", ErrBadHash)
}
}
return entry, fmt.Errorf("bad hash: could not place entry after %d resizes", defaultGrowthLimit)
return fmt.Errorf("could not place entry after %d resizes: %w", defaultGrowthLimit, ErrBadHash)
}
// Drop removes a value for a key in the table. Returns whether the key had
// existed.
func (t *Table[K, V]) Drop(key K) bool {
occupied := t.bucketA.drop(key) || t.bucketB.drop(key)
occupied := t.tableA.drop(key) || t.tableB.drop(key)
if t.load() < t.minLoadFactor {
// The error is not handled here, because table-shrinking is an internal
@@ -186,19 +198,19 @@ func (t *Table[K, V]) Drop(key K) bool {
}
// Entries returns an unordered sequence of all key-value pairs in the table.
func (t Table[K, V]) Entries() iter.Seq2[K, V] {
func (t *Table[K, V]) Entries() iter.Seq2[K, V] {
return func(yield func(K, V) bool) {
for _, slot := range t.bucketA.slots {
for _, slot := range t.tableA.slots {
if slot.occupied {
if !yield(slot.Key, slot.Value) {
if !yield(slot.key, slot.value) {
return
}
}
}
for _, slot := range t.bucketB.slots {
for _, slot := range t.tableB.slots {
if slot.occupied {
if !yield(slot.Key, slot.Value) {
if !yield(slot.key, slot.value) {
return
}
}
@@ -207,8 +219,8 @@ func (t Table[K, V]) Entries() iter.Seq2[K, V] {
}
// String returns the entries of the table as a string in the format:
// "table[k1:v1 h2:v2 ...]".
func (t Table[K, V]) String() string {
// "table[k1:v1 k2:v2 ...]".
func (t *Table[K, V]) String() string {
var sb strings.Builder
sb.WriteString("table[")
@@ -226,9 +238,9 @@ func (t Table[K, V]) String() string {
return sb.String()
}
// NewCustomTable creates a [Table] with custom [Hash] and [EqualFunc]
// NewCustom creates a [Table] with custom [Hash] and [EqualFunc]
// functions, along with any [Option] the user provides.
func NewCustomTable[K, V any](hashA, hashB Hash[K], compare EqualFunc[K], options ...Option) *Table[K, V] {
func NewCustom[K, V any](hashA, hashB Hash[K], compare EqualFunc[K], options ...Option) *Table[K, V] {
settings := &settings{
growthFactor: DefaultGrowthFactor,
bucketSize: DefaultCapacity,
@@ -242,8 +254,8 @@ func NewCustomTable[K, V any](hashA, hashB Hash[K], compare EqualFunc[K], option
return &Table[K, V]{
growthFactor: settings.growthFactor,
minLoadFactor: settings.minLoadFactor,
bucketA: newBucket[K, V](settings.bucketSize, hashA, compare),
bucketB: newBucket[K, V](settings.bucketSize, hashB, compare),
tableA: newSubtable[K, V](settings.bucketSize, hashA, compare),
tableB: newSubtable[K, V](settings.bucketSize, hashB, compare),
}
}
@@ -251,10 +263,10 @@ func pipe[X, Y, Z any](a func(X) Y, b func(Y) Z) func(X) Z {
return func(x X) Z { return b(a(x)) }
}
// NewTableBy creates a [Table] for any key type by using keyFunc to derive a
// NewBy creates a [Table] for any key type by using keyFunc to derive a
// comparable key. Two keys with the same derived key are treated as equal.
func NewTableBy[K, V any, C comparable](keyFunc func(K) C, options ...Option) *Table[K, V] {
return NewCustomTable[K, V](
func NewBy[K, V any, C comparable](keyFunc func(K) C, options ...Option) *Table[K, V] {
return NewCustom[K, V](
pipe(keyFunc, NewDefaultHash[C]()),
pipe(keyFunc, NewDefaultHash[C]()),
func(a, b K) bool { return keyFunc(a) == keyFunc(b) },
@@ -262,10 +274,10 @@ func NewTableBy[K, V any, C comparable](keyFunc func(K) C, options ...Option) *T
)
}
// NewTable creates a [Table] using the default [Hash] and [EqualFunc]. Use
// New creates a [Table] using the default [Hash] and [EqualFunc]. Use
// the [Option] functions to configure its behavior. Note that this constructor
// is only provided for comparable keys. For arbitrary keys, consider
// [NewTableBy] or [NewCustomTable].
func NewTable[K comparable, V any](options ...Option) *Table[K, V] {
return NewCustomTable[K, V](NewDefaultHash[K](), NewDefaultHash[K](), DefaultEqualFunc[K], options...)
// [NewBy] or [NewCustom].
func New[K comparable, V any](options ...Option) *Table[K, V] {
return NewCustom[K, V](NewDefaultHash[K](), NewDefaultHash[K](), DefaultEqualFunc[K], options...)
}