An efficient B-tree implementation in Go.
- Support for Generics (Go 1.18 ).
Map
andSet
types for ordered key-value maps and sets,- Fast bulk loading for pre-ordered data using the
Load()
method. Copy()
method with copy-on-write support.- Thread-safe operations.
- Path hinting optimization for operations with nearby keys.
To start using this package, install Go and run:
$ go get github.com/tidwall/btree
This package includes the following types of B-trees:
btree.Map
: A fast B-tree for storing ordered key value pairs. Go 1.18btree.Set
: LikeMap
, but only for storing keys. Go 1.18btree.Generic
: A feature-rich B-tree for storing data using a custom comparator. Go 1.18btree.BTree
: LikeGeneric
but uses theinterface{}
type for data. Backwards compatible. Go 1.16
// Basic
Set(key, value) // insert or replace an item
Get(key, value) // get an existing item
Delete(key) // delete an item
Len() // return the number of items in the map
// Iteration
Scan(iter) // scan items in ascending order
Reverse(iter) // scan items in descending order
Ascend(key, iter) // scan items in ascending order that are >= to key
Descend(key, iter) // scan items in descending order that are <= to key.
Iter() // returns a read-only iterator for for-loops.
// Array-like operations
GetAt(index) // returns the item at index
DeleteAt(index) // deletes the item at index
// Bulk-loading
Load(key, value) // load presorted items into tree
package main
import (
"fmt"
"github.com/tidwall/btree"
)
func main() {
// create a map
var users btree.Map[string, string]
// add some users
users.Set("user:4", "Andrea")
users.Set("user:6", "Andy")
users.Set("user:2", "Andy")
users.Set("user:1", "Jane")
users.Set("user:5", "Janet")
users.Set("user:3", "Steve")
// Iterate over the maps and print each user
users.Scan(func(key, value string) bool {
fmt.Printf("%s %s\n", key, value)
return true
})
fmt.Printf("\n")
// Delete a couple
users.Delete("user:5")
users.Delete("user:1")
// print the map again
users.Scan(func(key, value string) bool {
fmt.Printf("%s %s\n", key, value)
return true
})
fmt.Printf("\n")
// Output:
// user:1 Jane
// user:2 Andy
// user:3 Steve
// user:4 Andrea
// user:5 Janet
// user:6 Andy
//
// user:2 Andy
// user:3 Steve
// user:4 Andrea
// user:6 Andy
}
// Basic
Insert(key) // insert an item
Contains(key) // test if item exists
Delete(key) // delete an item
Len() // return the number of items in the set
// Iteration
Scan(iter) // scan items in ascending order
Reverse(iter) // scan items in descending order
Ascend(key, iter) // scan items in ascending order that are >= to key
Descend(key, iter) // scan items in descending order that are <= to key.
Iter() // returns a read-only iterator for for-loops.
// Array-like operations
GetAt(index) // returns the item at index
DeleteAt(index) // deletes the item at index
// Bulk-loading
Load(key) // load presorted item into tree
package main
import (
"fmt"
"github.com/tidwall/btree"
)
func main() {
// create a set
var names btree.Set[string]
// add some names
names.Insert("Jane")
names.Insert("Andrea")
names.Insert("Steve")
names.Insert("Andy")
names.Insert("Janet")
names.Insert("Andy")
// Iterate over the maps and print each user
names.Scan(func(key string) bool {
fmt.Printf("%s\n", key)
return true
})
fmt.Printf("\n")
// Delete a couple
names.Delete("Steve")
names.Delete("Andy")
// print the map again
names.Scan(func(key string) bool {
fmt.Printf("%s\n", key)
return true
})
fmt.Printf("\n")
// Output:
// Andrea
// Andy
// Jane
// Janet
// Steve
//
// Andrea
// Jane
// Janet
}
// Basic
Set(item) // insert or replace an item
Get(item) // get an existing item
Delete(item) // delete an item
Len() // return the number of items in the btree
// Iteration
Scan(iter) // scan items in ascending order
Reverse(iter) // scan items in descending order
Ascend(key, iter) // scan items in ascending order that are >= to key
Descend(key, iter) // scan items in descending order that are <= to key.
Iter() // returns a read-only iterator for for-loops.
// Array-like operations
GetAt(index) // returns the item at index
DeleteAt(index) // deletes the item at index
// Bulk-loading
Load(item) // load presorted items into tree
// Path hinting
SetHint(item, *hint) // insert or replace an existing item
GetHint(item, *hint) // get an existing item
DeleteHint(item, *hint) // delete an item
// Copy-on-write
Copy() // copy the btree
package main
import (
"fmt"
"github.com/tidwall/btree"
)
type Item struct {
Key, Val string
}
// byKeys is a comparison function that compares item keys and returns true
// when a is less than b.
func byKeys(a, b Item) bool {
return a.Key < b.Key
}
// byVals is a comparison function that compares item values and returns true
// when a is less than b.
func byVals(a, b Item) bool {
if a.Val < b.Val {
return true
}
if a.Val > b.Val {
return false
}
// Both vals are equal so we should fall though
// and let the key comparison take over.
return byKeys(a, b)
}
func main() {
// Create a tree for keys and a tree for values.
// The "keys" tree will be sorted on the Keys field.
// The "values" tree will be sorted on the Values field.
keys := btree.NewGeneric[Item](byKeys)
vals := btree.NewGeneric[Item](byVals)
// Create some items.
users := []Item{
Item{Key: "user:1", Val: "Jane"},
Item{Key: "user:2", Val: "Andy"},
Item{Key: "user:3", Val: "Steve"},
Item{Key: "user:4", Val: "Andrea"},
Item{Key: "user:5", Val: "Janet"},
Item{Key: "user:6", Val: "Andy"},
}
// Insert each user into both trees
for _, user := range users {
keys.Set(user)
vals.Set(user)
}
// Iterate over each user in the key tree
keys.Scan(func(item Item) bool {
fmt.Printf("%s %s\n", item.Key, item.Val)
return true
})
fmt.Printf("\n")
// Iterate over each user in the val tree
vals.Scan(func(item Item) bool {
fmt.Printf("%s %s\n", item.Key, item.Val)
return true
})
// Output:
// user:1 Jane
// user:2 Andy
// user:3 Steve
// user:4 Andrea
// user:5 Janet
// user:6 Andy
//
// user:4 Andrea
// user:2 Andy
// user:6 Andy
// user:1 Jane
// user:5 Janet
// user:3 Steve
}
// Basic
Set(item) // insert or replace an item
Get(item) // get an existing item
Delete(item) // delete an item
Len() // return the number of items in the btree
// Iteration
Scan(iter) // scan items in ascending order
Reverse(iter) // scan items in descending order
Ascend(key, iter) // scan items in ascending order that are >= to key
Descend(key, iter) // scan items in descending order that are <= to key.
Iter() // returns a read-only iterator for for-loops.
// Array-like operations
GetAt(index) // returns the item at index
DeleteAt(index) // deletes the item at index
// Bulk-loading
Load(item) // load presorted items into tree
// Path hinting
SetHint(item, *hint) // insert or replace an existing item
GetHint(item, *hint) // get an existing item
DeleteHint(item, *hint) // delete an item
// Copy-on-write
Copy() // copy the btree
package main
import (
"fmt"
"github.com/tidwall/btree"
)
type Item struct {
Key, Val string
}
// byKeys is a comparison function that compares item keys and returns true
// when a is less than b.
func byKeys(a, b interface{}) bool {
i1, i2 := a.(*Item), b.(*Item)
return i1.Key < i2.Key
}
// byVals is a comparison function that compares item values and returns true
// when a is less than b.
func byVals(a, b interface{}) bool {
i1, i2 := a.(*Item), b.(*Item)
if i1.Val < i2.Val {
return true
}
if i1.Val > i2.Val {
return false
}
// Both vals are equal so we should fall though
// and let the key comparison take over.
return byKeys(a, b)
}
func main() {
// Create a tree for keys and a tree for values.
// The "keys" tree will be sorted on the Keys field.
// The "values" tree will be sorted on the Values field.
keys := btree.New(byKeys)
vals := btree.New(byVals)
// Create some items.
users := []*Item{
&Item{Key: "user:1", Val: "Jane"},
&Item{Key: "user:2", Val: "Andy"},
&Item{Key: "user:3", Val: "Steve"},
&Item{Key: "user:4", Val: "Andrea"},
&Item{Key: "user:5", Val: "Janet"},
&Item{Key: "user:6", Val: "Andy"},
}
// Insert each user into both trees
for _, user := range users {
keys.Set(user)
vals.Set(user)
}
// Iterate over each user in the key tree
keys.Ascend(nil, func(item interface{}) bool {
kvi := item.(*Item)
fmt.Printf("%s %s\n", kvi.Key, kvi.Val)
return true
})
fmt.Printf("\n")
// Iterate over each user in the val tree
vals.Ascend(nil, func(item interface{}) bool {
kvi := item.(*Item)
fmt.Printf("%s %s\n", kvi.Key, kvi.Val)
return true
})
// Output:
// user:1 Jane
// user:2 Andy
// user:3 Steve
// user:4 Andrea
// user:5 Janet
// user:6 Andy
//
// user:4 Andrea
// user:2 Andy
// user:6 Andy
// user:1 Jane
// user:5 Janet
// user:3 Steve
}
This implementation was designed with performance in mind.
google
: The google/btree package (without generics)tidwall
: The tidwall/btree package (without generics)tidwall(G)
: The tidwall/btree package (generics using thebtree.Generic
type)tidwall(M)
: The tidwall/btree package (generics using thebtree.Map
type)go-arr
: A simple Go array
The following benchmarks were run on my 2019 Macbook Pro (2.4 GHz 8-Core Intel Core i9) using Go Development version 1.18 (beta1). The items are simple 8-byte ints.
** sequential set **
google: set-seq 1,000,000 ops in 156ms, 6,426,724/sec, 155 ns/op, 39.0 MB, 40 bytes/op
tidwall: set-seq 1,000,000 ops in 135ms, 7,380,627/sec, 135 ns/op, 23.5 MB, 24 bytes/op
tidwall(G): set-seq 1,000,000 ops in 78ms, 12,881,995/sec, 77 ns/op, 8.2 MB, 8 bytes/op
tidwall(M): set-seq 1,000,000 ops in 46ms, 21,892,141/sec, 45 ns/op, 8.2 MB, 8 bytes/op
tidwall: set-seq-hint 1,000,000 ops in 73ms, 13,789,017/sec, 72 ns/op, 23.5 MB, 24 bytes/op
tidwall(G): set-seq-hint 1,000,000 ops in 48ms, 20,969,431/sec, 47 ns/op, 8.2 MB, 8 bytes/op
tidwall: load-seq 1,000,000 ops in 45ms, 22,452,523/sec, 44 ns/op, 23.5 MB, 24 bytes/op
tidwall(G): load-seq 1,000,000 ops in 22ms, 46,242,274/sec, 21 ns/op, 8.2 MB, 8 bytes/op
tidwall(M): load-seq 1,000,000 ops in 13ms, 74,371,903/sec, 13 ns/op, 8.2 MB, 8 bytes/op
go-arr: append 1,000,000 ops in 21ms, 47,141,875/sec, 21 ns/op, 8.1 MB, 8 bytes/op
** sequential get **
google: get-seq 1,000,000 ops in 119ms, 8,389,459/sec, 119 ns/op
tidwall: get-seq 1,000,000 ops in 110ms, 9,068,759/sec, 110 ns/op
tidwall(G): get-seq 1,000,000 ops in 78ms, 12,813,135/sec, 78 ns/op
tidwall(M): get-seq 1,000,000 ops in 62ms, 16,053,728/sec, 62 ns/op
tidwall: get-seq-hint 1,000,000 ops in 64ms, 15,509,696/sec, 64 ns/op
tidwall(G): get-seq-hint 1,000,000 ops in 41ms, 24,144,951/sec, 41 ns/op
** random set **
google: set-rand 1,000,000 ops in 563ms, 1,777,592/sec, 562 ns/op, 29.7 MB, 31 bytes/op
tidwall: set-rand 1,000,000 ops in 542ms, 1,844,397/sec, 542 ns/op, 29.6 MB, 31 bytes/op
tidwall(G): set-rand 1,000,000 ops in 234ms, 4,271,764/sec, 234 ns/op, 11.2 MB, 11 bytes/op
tidwall(M): set-rand 1,000,000 ops in 189ms, 5,292,236/sec, 188 ns/op, 11.2 MB, 11 bytes/op
tidwall: set-rand-hint 1,000,000 ops in 602ms, 1,659,852/sec, 602 ns/op, 29.6 MB, 31 bytes/op
tidwall(G): set-rand-hint 1,000,000 ops in 278ms, 3,595,435/sec, 278 ns/op, 11.2 MB, 11 bytes/op
tidwall: set-after-copy 1,000,000 ops in 679ms, 1,471,954/sec, 679 ns/op
tidwall(G): set-after-copy 1,000,000 ops in 238ms, 4,196,854/sec, 238 ns/op
tidwall: load-rand 1,000,000 ops in 532ms, 1,880,877/sec, 531 ns/op, 29.6 MB, 31 bytes/op
tidwall(G): load-rand 1,000,000 ops in 232ms, 4,316,475/sec, 231 ns/op, 11.2 MB, 11 bytes/op
tidwall(M): load-rand 1,000,000 ops in 209ms, 4,790,169/sec, 208 ns/op, 11.2 MB, 11 bytes/op
** random get **
google: get-rand 1,000,000 ops in 807ms, 1,238,703/sec, 807 ns/op
tidwall: get-rand 1,000,000 ops in 812ms, 1,231,551/sec, 811 ns/op
tidwall(G): get-rand 1,000,000 ops in 255ms, 3,914,819/sec, 255 ns/op
tidwall(M): get-rand 1,000,000 ops in 190ms, 5,249,966/sec, 190 ns/op
tidwall: get-rand-hint 1,000,000 ops in 876ms, 1,141,313/sec, 876 ns/op
tidwall(G): get-rand-hint 1,000,000 ops in 258ms, 3,877,775/sec, 257 ns/op
** range **
google: ascend 1,000,000 ops in 26ms, 39,101,882/sec, 25 ns/op
tidwall: ascend 1,000,000 ops in 20ms, 50,223,988/sec, 19 ns/op
tidwall(G): iter 1,000,000 ops in 8ms, 119,155,937/sec, 8 ns/op
tidwall(G): scan 1,000,000 ops in 6ms, 168,275,407/sec, 5 ns/op
tidwall(G): walk 1,000,000 ops in 5ms, 186,941,046/sec, 5 ns/op
go-arr: for-loop 1,000,000 ops in 4ms, 272,234,997/sec, 3 ns/op
You can find the benchmark utility at tidwall/btree-benchmark
Josh Baker @tidwall
Source code is available under the MIT License.