From the course: Faster Python Code

Unlock the full course today

Join today to access over 22,600 courses taught by industry experts or purchase this course individually.

When approximation is good enough

When approximation is good enough - Python Tutorial

From the course: Faster Python Code

Start my 1-month free trial

When approximation is good enough

“

- [Instructor] Here is a riddle for you. How much is 1.1 times 1.1? Let's see what Python says. So we'll do 1.1 times 1.1. I find the bargain python posts from time to time. A lot of them are similar to what you just saw. Usually applies to RDM failing. Which stands for redefine manual. We send them to read about floating point implementation. It is not the bargain python, if you run the same code in C, Java, Go and other languages, you see the same results. The reason is a design creative. To make floating point calculations fast, some degree of accuracy is sacrificed. Look at the answer, the error is very small. But still, it is an error. As a side note, every time you do testing that involves checking results of floating point numbers, you should allow for some error. Use the built-in round function with the number of digits you think is an acceptable error. What I'm trying to show you that even in something as mainstream as floating points, using approximation or we cheat. The…

Contents