Floating point types
Floating point types allow you to represent real numbers, not just integers. This allows you to work with a continuous and theoretically infinite set of values within your computations. It may come as no surprise that floating point calculations show up in almost every scientific computation, macro-financial analysis, machine learning algorithm, and so on.
The emphasis on the word theoretically, however, is intentional and very important to understand. Floating point types still have boundaries, with real limitations being imposed by your computer hardware. In essence, the notion of being able to represent any number is an illusion. Floating point types are liable to lose precision and introduce rounding errors, especially as you work with more extreme values. As such, floating point types are not suitable when you need absolute precision (for that, you will want to reference the PyArrow decimal types introduced later in this chapter).
Despite those limitations...
