提升性能的utf8.DecodeRuneInString

普通的截取

var benchmarkStr = "此次进行的是中文的截取的性能测试情况"


func StrSplit(i,j int,str string)string{
	return str[i:j]
}

func BenchmarkStrSplit(b *testing.B) {
	for i := 0; i < b.N; i++ {
		StrSplit(3,10, benchmarkStr)
	}
}

结果

goos: darwin
goarch: amd64
pkg: test/string/src
BenchmarkStrSplit-4     1000000000               0.579 ns/op           0 B/op          0 allocs/op
PASS
ok      test/string/src 0.648s
go test -bench=StrSplit -benchmem  1.00s user 0.32s system 119% cpu 1.107 total

可以看出速度是可以的 ，也没有内存的分配申请，但是中文会产生乱码

中问的截取[]rune

var benchmarkStr = "此次进行的是中文的截取的性能测试情况"

func SplitStrRunes(s string, length int) string {
	if utf8.RuneCountInString(s) > length {
		rs := []rune(s)
		return string(rs[:length])
	}

	return s
}

func BenchmarkSubStrRunes(b *testing.B) {
	for i := 0; i < b.N; i++ {
		SplitStrRunes(benchmarkStr, 10)
	}
}

goos: darwin
goarch: amd64
pkg: test/string/src
BenchmarkStrSplit-4             1000000000               0.573 ns/op           0 B/op          0 allocs/op
BenchmarkSubStrRunes-4           2704357               443 ns/op              48 B/op          1 allocs/op
PASS
ok      test/string/src 2.288s
go test -bench=. -benchmem  2.63s user 0.22s system 110% cpu 2.569 total

时间相比传统的截取，发生了内存拷贝，并且时间很长

提升性能的utf8.DecodeRuneInString

func BenchmarkStrSplit(b *testing.B) {
	for i := 0; i < b.N; i++ {
		StrSplit(3,10, benchmarkStr)
	}
}


func SubStrDecodeRuneInString(s string, length int) string {
	var size, n int
	for i := 0; i < length && n < len(s); i++ {
		_, size = utf8.DecodeRuneInString(s[n:])
		n += size
	}

	return s[:n]
}

结果

goos: darwin
goarch: amd64
pkg: test/string/src
BenchmarkStrSplit-4                     1000000000               0.620 ns/op           0 B/op          0 allocs/op
BenchmarkSubStrDecodeRuneInString-4     18224007                61.1 ns/op             0 B/op          0 allocs/op
BenchmarkSubStrRunes-4                   2541043               439 ns/op              48 B/op          1 allocs/op
PASS
ok      test/string/src 3.470s
go test -bench=. -benchmem  3.83s user 0.52s system 101% cpu 4.298 total

结果和[]rune相比性能提升了很多，而且减少了内存分配的次数，但是和传统饮用还是有很大的差距

for range迭代字符，不是字节

func SubStrRange(s string, length int) string {
	var n, i int
	for i = range s {
		if n == length {
			break
		}

		n++
	}

	return s[:i]
}

func BenchmarkSubStrRange(b *testing.B) {
	for i := 0; i < b.N; i++ {
		SubStrRange(benchmarkStr, 10)
	}
}

结果

goos: darwin
goarch: amd64
pkg: test/string/src
BenchmarkStrSplit-4                     1000000000               0.582 ns/op           0 B/op          0 allocs/op
BenchmarkSubStrDecodeRuneInString-4     19436258                61.8 ns/op             0 B/op          0 allocs/op
BenchmarkSubStrRunes-4                   2709249               454 ns/op              48 B/op          1 allocs/op
BenchmarkSubStrRange-4                  13656135                87.3 ns/op             0 B/op          0 allocs/op
PASS
ok      test/string/src 4.877s
go test -bench=. -benchmem  5.23s user 0.58s system 101% cpu 5.733 total

多次结果发现 for _range 是和utf8.DecodeRuneInString 差不多的，但是要稍微慢点 还是原始的最快