in reply to Re: Question about regex performance (too small)
in thread Question about regex performance
You are absolutely right. The fact is that if I were working on improving the speed of production code, I would begin with profiling. In THAT activity, there is a balance of costs between how quickly code can be developed versus how fast the code can be. I have seen cases where the fastest code is extremely hard for junior or even intermediate coders to understand. Sometime it is necessary to implement that anyway because the cost in lost time due to slower algorithms is much greater. But sometimes, the simpler, slower code is desired because it can be implemented quickly by your least experienced staff.
Alas, you missed the point of the exercise. My objective is to understand these regular expressions better. I rarely develop them, and when I do, I must have the manual for regular expressions open, in order to figure out how to develop one that meets my needs. I do not need such assistance when I deal with solving systems of linear equations, numeric quadrature, or statistical analysis. I am, in a sense, pushing myself out of my comfort zone. In this context, the benchmark scripts I show are mere devices to provide one way of evaluating the merits of the different algorithms I found or developed. And, while I do intend to modify them to work with longer strings, what I was especially hoping for is some insight into the reasons for the difference in performance and how to combine the regular expressions that trim leading and trailing white space with those that eliminate redundant white space characters within strings; or if such a combination even makes sense.