Abstract
Polars boasts 18 different data types, not including variants of numerical types.
Do we really need such a vast collection of data types?
What is the use case for each type?
What is the difference between List and Array? Or between Categorical and Enum? And why on Earth would I ever need a Struct?
This talk will clear up all of these questions and more, as we go through the data types that Polars provides and understand why we need each one of them.
Description
According to the documentation, Polars has 18 data types (excluding the varying precision of numerical data types).
The use cases for some data types are intuitively very clear. For example, we all know when to use Booleans, integers, floating-point numbers, or strings.
Some pairs of data types are fairly easy to understand, but their distinctions can be fuzzy.
For example, when do you use List or Array?
When is Categorical better than Enum and vice-versa?
And some less common data types are poorly understood, like Decimal, Object, or Struct.
Watch the recording
