GO tree

AVL tree implementation

• very fast insertion, look ups, and deletions
• very fast iteration

Basic Usage

import (
"fmt"
"github.com/ross-oreto/go-tree"
)

btree := tree.New()
btree.Insert(StringVal("Oreto")).Insert(StringVal("Michael")).Insert(StringVal("Ross"))
fmt.Println(btree.Values())

Retrieve

btree.Get(StringVal("Ross"))

Delete

btree.Delete(StringVal("Ross"))

Clear

btree.Init()

Val type

Values entered into the tree must implement the Val interface Comp method

• Comp(val Val) int8

type TestKey1 struct {
	Name string
}
// Comp returns 1 if key > val, -1 if key < val and 0 if key equal to val
func (key TestKey1) Comp(val Val) int8 {
	var c int8
	tk := val.(TestKey1)
	if key.Name > tk.Name {
		c = 1
	} else if key.Name < tk.Name {
		c = -1
	}
	return c
}

Performance

Benchmarks ran against "github.com/google/btree" using a random tree with 1 million inserts

BenchmarkBtree_Insert-4                        3         357906766 ns/op        24000016 B/op    1333333 allocs/op
BenchmarkBtree_Insert-4                        3         356572333 ns/op        24000016 B/op    1333333 allocs/op
BenchmarkGtree_Insert-4                        3         364911600 ns/op        25461704 B/op    1032939 allocs/op
BenchmarkGtree_Insert-4                        3         366578900 ns/op        25461698 B/op    1032939 allocs/op
BenchmarkBtree_InsertRandom-4                  1        1354908200 ns/op        56000032 B/op    2000000 allocs/op
BenchmarkBtree_InsertRandom-4                  1        1351906700 ns/op        56000000 B/op    2000000 allocs/op
BenchmarkGtree_InsertRandom-4                  1        1049704500 ns/op        44105176 B/op    1068918 allocs/op
BenchmarkGtree_InsertRandom-4                  1        1070716500 ns/op        44105176 B/op    1068918 allocs/op
BenchmarkBtree_Get-4                           2         984660150 ns/op         8000000 B/op    1000000 allocs/op
BenchmarkBtree_Get-4                           2         969647750 ns/op         8000000 B/op    1000000 allocs/op
BenchmarkGtree_Get-4                           1        1065713900 ns/op         8000000 B/op    1000000 allocs/op
BenchmarkGtree_Get-4                           1        1064712600 ns/op         8000000 B/op    1000000 allocs/op
BenchmarkBtree_Iteration-4                  2000            591395 ns/op               0 B/op          0 allocs/op
BenchmarkBtree_Iteration-4                  2000            590395 ns/op               0 B/op          0 allocs/op
BenchmarkGtree_Iteration-4                   100          10356963 ns/op               0 B/op          0 allocs/op
BenchmarkGtree_Iteration-4                   100          10346943 ns/op               0 B/op          0 allocs/op
BenchmarkBtree_Len-4                    2000000000               0.59 ns/op            0 B/op          0 allocs/op
BenchmarkBtree_Len-4                    2000000000               0.59 ns/op            0 B/op          0 allocs/op
BenchmarkGtree_Len-4                    2000000000               0.59 ns/op            0 B/op          0 allocs/op
BenchmarkGtree_Len-4                    2000000000               0.59 ns/op            0 B/op          0 allocs/op
BenchmarkBtree_Delete-4                       10         122482200 ns/op         6400001 B/op     300000 allocs/op
BenchmarkBtree_Delete-4                       10         121581680 ns/op         6400001 B/op     300000 allocs/op
BenchmarkGtree_Delete-4                       10         115377360 ns/op         5217883 B/op     206879 allocs/op
BenchmarkGtree_Delete-4                       10         113776200 ns/op         5217884 B/op     206879 allocs/op

• btree = go-tree
• gtree = google/btree

Results show:

• Insertions are faster with btree when the list is in order however when random gtree performs faster
• Gets are faster using btree
• Iteration much faster with btree
• Length reporting at the same speed
• tree deletions are slightly faster using gtree
• A larger memory footprint using btree during insertions and deletions