Python for Finance, 2nd Edition

Chapter 4. Numerical Computing with NumPy

Computers are useless. They can only give answers.

Pablo Picasso

Although the Python interpreter itself already brings a rich variety of data structures with it, NumPy and other libraries add to these in a valuable fashion. This chapter focuses on NumPy, which provides a multidimensional array object to store homogeneous or heterogeneous data arrays and supports vectorization of code.

The chapter covers the following data structures:

Object type	Meaning	Used for
ndarray (regular)	n-dimensional array object	Large arrays of numerical data
ndarray (record)	2-dimensional array object	Tabular data organized in columns

This chapter is organized as follows:

“Arrays of Data”

This section is about the handling of arrays of data with pure Python code.

“Regular NumPy Arrays”

This is the core section about the regular NumPy ndarray class, the workhorse in almost all data-intensive Python use cases involving numerical data.

“Structured NumPy Arrays”

This brief section introduces structured (or record) ndarray objects for the handling of tabular data with columns.

“Vectorization of Code”

In this section, vectorization of code is discussed along with its benefits; the section also discusses the importance of memory layout in certain scenarios.

Arrays of Data

The previous chapter showed that Python provides some quite useful and flexible general data structures. In particular, list objects can be considered a real workhorse with ...