C-Sharp | Java | Python | Swift | GO | WPF | Ruby | Scala | F# | JavaScript | SQL | PHP | Angular | HTML
However, it does not eliminate the cost of individual appends. This benchmark shows that performance improves by appending fewer, larger strings. Appending many smaller strings takes longer overall.
Benchmark. Let's look at how the data was collected. The goal was to create a StringBuilder with five million characters in it. We append individual strings of lengths 1 to 99 characters to reach that final length. The Append method calls are timed.
C# program that times StringBuilder appends using System; using System.Diagnostics; using System.Text; class Program { const int _length = 5000000; const int _max = 1000000; static void Main() { StringBuilder builder = new StringBuilder(50000 + 100); for (int a = 1; a < 100; a++) { string v = new string('x', a); int repeat = _length / a; var s1 = Stopwatch.StartNew(); for (int i = 0; i < repeat; i++) { builder.Append(v); } s1.Stop(); builder.Clear(); GC.Collect(); Console.WriteLine("{0} chars x {1}: {2:0.00 ms}", a, repeat, ((double)(s1.Elapsed.TotalMilliseconds))); } Console.WriteLine("[Done]"); Console.Read(); } } Output 1 chars x 5000000: 47.02 ms 2 chars x 2500000: 21.26 ms 3 chars x 1666666: 16.93 ms 4 chars x 1250000: 12.35 ms 5 chars x 1000000: 10.72 ms 6 chars x 833333: 9.92 ms 7 chars x 714285: 8.20 ms 8 chars x 625000: 6.80 ms 9 chars x 555555: 8.24 ms 10 chars x 500000: 6.06 ms 11 chars x 454545: 6.54 ms 12 chars x 416666: 5.29 ms 13 chars x 384615: 6.43 ms 14 chars x 357142: 4.90 ms 15 chars x 333333: 5.77 ms 16 chars x 312500: 4.63 ms 17 chars x 294117: 4.58 ms 18 chars x 277777: 4.40 ms 19 chars x 263157: 4.32 ms 20 chars x 250000: 4.25 ms 21 chars x 238095: 4.61 ms 22 chars x 227272: 4.06 ms 23 chars x 217391: 4.74 ms 24 chars x 208333: 3.88 ms 25 chars x 200000: 4.30 ms 26 chars x 192307: 4.45 ms 27 chars x 185185: 4.40 ms 28 chars x 178571: 4.29 ms 29 chars x 172413: 9.04 ms 30 chars x 166666: 4.25 ms 31 chars x 161290: 4.41 ms 32 chars x 156250: 3.65 ms 33 chars x 151515: 4.34 ms 34 chars x 147058: 4.59 ms 35 chars x 142857: 4.40 ms 36 chars x 138888: 4.19 ms 37 chars x 135135: 4.92 ms 38 chars x 131578: 4.31 ms 39 chars x 128205: 4.49 ms 40 chars x 125000: 4.50 ms 41 chars x 121951: 4.36 ms 42 chars x 119047: 4.71 ms 43 chars x 116279: 4.03 ms 44 chars x 113636: 3.61 ms 45 chars x 111111: 4.30 ms 46 chars x 108695: 4.47 ms 47 chars x 106382: 4.54 ms 48 chars x 104166: 3.62 ms 49 chars x 102040: 4.10 ms 50 chars x 100000: 4.15 ms 51 chars x 98039: 4.10 ms 52 chars x 96153: 3.52 ms 53 chars x 94339: 4.10 ms 54 chars x 92592: 4.09 ms 55 chars x 90909: 4.27 ms 56 chars x 89285: 4.11 ms 57 chars x 87719: 4.20 ms 58 chars x 86206: 4.23 ms 59 chars x 84745: 4.25 ms 60 chars x 83333: 4.44 ms 61 chars x 81967: 4.22 ms 62 chars x 80645: 4.26 ms 63 chars x 79365: 4.26 ms 64 chars x 78125: 4.22 ms 65 chars x 76923: 4.22 ms 66 chars x 75757: 4.27 ms 67 chars x 74626: 4.20 ms 68 chars x 73529: 4.32 ms 69 chars x 72463: 3.73 ms 70 chars x 71428: 4.18 ms 71 chars x 70422: 4.48 ms 72 chars x 69444: 3.66 ms 73 chars x 68493: 4.28 ms 74 chars x 67567: 3.77 ms 75 chars x 66666: 4.11 ms 76 chars x 65789: 4.31 ms 77 chars x 64935: 3.82 ms 78 chars x 64102: 4.30 ms 79 chars x 63291: 3.77 ms 80 chars x 62500: 4.19 ms 81 chars x 61728: 3.92 ms 82 chars x 60975: 4.76 ms 83 chars x 60240: 4.21 ms 84 chars x 59523: 4.08 ms 85 chars x 58823: 4.32 ms 86 chars x 58139: 4.30 ms 87 chars x 57471: 4.40 ms 88 chars x 56818: 4.59 ms 89 chars x 56179: 4.30 ms 90 chars x 55555: 4.40 ms 91 chars x 54945: 4.26 ms 92 chars x 54347: 3.86 ms 93 chars x 53763: 4.38 ms 94 chars x 53191: 3.54 ms 95 chars x 52631: 3.75 ms 96 chars x 52083: 4.45 ms 97 chars x 51546: 4.13 ms 98 chars x 51020: 4.20 ms 99 chars x 50505: 4.48 ms
Discussion. How does this benchmark help in real programs? If you have a program that calls Append on strings with fewer than about 20 characters, it could be made faster by using 20 or more characters.
Also: There is no benefit to using more than 20 characters at a time. Combining Append calls could make a performance difference.
Summary. Programs that use StringBuilder will perform faster if some calls to Append are combined if possible. Coalescing Append calls could lead to a performance gain. However, this only helps when appending strings less than about 20 characters.