字符串拼接性能对比

发表于 | 分类于

字符串拼接性能对比

笔者在日常工作中经常会遇到将数据进行批处理，即接收多个数据源发送的数据后合并批量操作。在一次针对字符串拼接时出现处理耗时达到 1s（偶发情况），在排查问题时发现在高并发的情况下由于使用了 + 直接对字符串进行拼接导致系统处理时间很慢，因此进行优化。

下面基于常用的字符串拼接进行性能对比：

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
func BenchmarkSprintf(b *testing.B) {
    for i := 0; i < b.N; i++ {
       var values string
       for _, s := range bufData {
          if s != "" {
             values += fmt.Sprintf("(%s),", s)
          }
       }
       _ = strings.TrimRight(values, ",")
    }
}

func BenchmarkBuilder(b *testing.B) {
    for i := 0; i < b.N; i++ {
       var buf strings.Builder
       for _, s := range bufData {
          if s != "" {
             buf.WriteString("(")
             buf.WriteString(s)
             buf.WriteString(")")
             buf.WriteString(",")
          }
       }
       _ = strings.TrimRight(buf.String(), ",")
    }
}

func BenchmarkBufferWriteString(b *testing.B) {
    for i := 0; i < b.N; i++ {
       var buf bytes.Buffer
       for _, s := range bufData {
          if s != "" {
             buf.WriteString("(")
             buf.WriteString(s)
             buf.WriteString("),")
          }
       }
       _ = strings.TrimRight(buf.String(), ",")
    }
}

func BenchmarkBufferWriteByte(b *testing.B) {
    for i := 0; i < b.N; i++ {
       var buf bytes.Buffer
       for _, s := range bufData {
          if s != "" {
             buf.Write([]byte("("))
             buf.Write([]byte(s))
             buf.Write([]byte("),"))
          }
       }
       _ = strings.TrimRight(buf.String(), ",")
    }
}

总结

在高并发情况下，使用 bytes.Buffer 进行字符串拼接的效率要比使用 + 或者 Sprintf 的效率高出好几倍。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
$ go test -benchmem -run=^$ -count=3  -bench "^Benchmar.*$"
goos: windows
goarch: amd64
pkg: example/str
cpu: Intel(R) Core(TM) i5-9400 CPU @ 2.90GHz
BenchmarkSprintf-6                  7252            156178 ns/op         1243382 B/op        105 allocs/op
BenchmarkSprintf-6                  8191            160537 ns/op         1243474 B/op        105 allocs/op
BenchmarkSprintf-6                  8370            151687 ns/op         1243526 B/op        105 allocs/op
BenchmarkBuilder-6                 35635             33629 ns/op          273802 B/op         12 allocs/op
BenchmarkBuilder-6                 33496             35294 ns/op          273802 B/op         12 allocs/op
BenchmarkBuilder-6                 35260             35085 ns/op          273801 B/op         12 allocs/op
BenchmarkBufferWriteString-6       47410             26383 ns/op          194368 B/op          8 allocs/op
BenchmarkBufferWriteString-6       46074             26450 ns/op          194368 B/op          8 allocs/op
BenchmarkBufferWriteString-6       44607             25640 ns/op          194368 B/op          8 allocs/op
BenchmarkBufferWriteByte-6         35469             34121 ns/op          257089 B/op         43 allocs/op
BenchmarkBufferWriteByte-6         37026             36862 ns/op          257089 B/op         43 allocs/op
BenchmarkBufferWriteByte-6         34782             35819 ns/op          257089 B/op         43 allocs/op
PASS
ok      example/str     18.770s