I don't see how JoinArgs that I wrote is much slower since it uses the stack to push the arguments, and then String.Join probably just uses StringBuilder anyways I'd assume. I simply wrote it to replace instances where I used a ton of string concatenations using the + (plus) operator. It makes it so I can just find and replace all "+" (pluses) with "," (commas) and wrap the entire thing in JoinArgs ().
On Feb 5, 11:27 am, Cerebrus <[email protected]> wrote: > Everyone has expressed their opinions, maybe I should too! ;-) > > Although I have great respect for Mahesh Chand, I think that that > particular article is only for newbies. Using DateTime.Now().ToString > () is hardly a good way to evaluate time elapsed. Additionally, such > measurements should be in ticks, seconds are too large a unit to > detect minute performance differences that only show up over hundreds > of thousands of iterations. Thirdly, performance is usually gauged as > an average time over a number of test repetitions. (Say, you run the > code 20 times and calculate the average time taken) > > Finally, I always prefer to use the StringBuilder any time I am > concatenating more than 5 strings. When using the StringBuilder, I try > to guess an appropriate capacity(usually half or slightly more than > estimated total capacity) for the object and use the overload that > accepts a starting size. For instance, > > --- > // "example" has 7 characters, so 7 x 20 will be 140 chars. Since this > is a small number, I'll use 150 (round figure). > // Now the StringBuilder will not need to double its buffer at all. > StringBuilder sb = new StringBuilder(150) > for (int i = 0; i < 20; i++) > { > sb.Append("example");} > > --- > > Note that I am not saying that this code is noticeably faster than > concatenating 20 instances of a string. I only intend to demonstrate > the *best* way to use a StringBuilder. Since the default initial > capacity is only 16, it means that the StringBuilder would have to > reallocate buffers atleast 4 times to suffice for a 140 char capacity. > (16+32+64+128) > > Hope that clears up any persisting doubts. > > -- > Cerebrus. > > On Feb 5, 1:19 pm, Brandon Betances <[email protected]> wrote: > > > > >http://www.c-sharpcorner.com/UploadFile/mahesh/StringBuilderComp11232... > > > It's obvious he's not concatenating more than 10 strings. The proof is in > > the muri.- Hide quoted text - > > - Show quoted text -
