Generate test dataframe with polars (powered by hypothesis)

The sub-module polars.testing.parametric provide tools for generating fake data for testing purposes. Here is an example showing what can be done with just dataframes and column functions in this module

import polars as pl
from polars.testing.parametric import dataframes, column

def generate(size=5):
    return dataframes(
    [
        column("id", dtype=pl.UInt16, unique=True, allow_null=False), 
        column("value", dtype=pl.Int16, allow_null=True), 
        column("cat", dtype =pl.Enum("XYZ"), allow_null=False)
    ], 
    min_size=size, max_size=size)

original = generate().example()

The output is random, i.e. evey call to the example method would generate a new dataframe with the prescribed characteristics (this method is for interactive use only). One can test their data pipelines on fake data with precise schema and simulated data quality deficiencies (eg null values, nan, inf, etc).

For unittesting, here is an example from the offical docs

from hypothesis import given
@given(df=dataframes(min_size=3, max_size=5))
def test_df_height(df: pl.DataFrame) -> None:
    assert 3 <= df.height <= 5