Python for Finance, 2nd Edition

Chapter 10. Performance Python

Don’t lower your expectations to meet your performance. Raise your level of performance to meet your expectations.

Ralph Marston

It is a long-lived prejudice that Python per se is a relatively slow programming language and not appropriate to implement computationally demanding tasks in finance. Beyond the fact that Python is an interpreted language, the reasoning is usually along the following lines: Python is slow when it comes to loops; loops are often required to implement financial algorithms; therefore Python is too slow for financial algorithm implementation. Another line of reasoning is: other (compiled) programming languages are fast at executing loops (such as C or C++); loops are often required for financial algorithms; therefore these (compiled) programming languages are well suited for finance and financial algorithm implementation.

Admittedly, it is possible to write proper Python code that executes rather slowly—perhaps too slowly for many application areas. This chapter is about approaches to speed up typical tasks and algorithms often encountered in a financial context. It shows that with a judicious use of data structures, choosing the right implementation idioms and paradigms, as well as using the right performance packages, Python is able to compete even with compiled programming languages. This is due to, among other factors, getting compiled itself.

To this end, this chapter introduces different approaches to speed up code: ...