Sorry not helpful for your situation exactly, but I don’t know about regex being easier to read than an old C style method.
On Sat, 22 Jan 2022 at 11:49, Greg Keogh <gfke...@gmail.com> wrote: > Folks, I have to parse just over a million lines of text as quickly as > possible. I'm using a Regex instance with a simple pattern (no backtrack or > the like) and it's taking from 1 to 2 seconds, which my colleague says is > too slow. > > Since it's a .NET 6 library I was keen to use the Span<> overloads of > Regex.Match which should perform much better due to reduced allocations, > but I can't find them. I know there was a big release of Span overloads > recently, but didn't it make it into Regex? > > FYI -- My colleague used to parse the lines the old C-style way using a for > (int =0; i<line.Length; i++) and looking for delimiters and *eating* > substrings > as he went. That 1970s style code was hard to read, but it performed about > 10x faster than a Regex. I thought my demo with a Regex.Match and just a > few lines of code would be faster, but now I have egg on my face. That's > why I was looking for the Span overloads to save my reputation. > > *Greg K* >
