Create hierarchy on text data in polars

Enum

Polars has a special data type that allows you to filter text values based on hierarchy. Let’s create a dataframe where the Degree column is of the data type Enum.

import polars as pl

education_level = ["High school", "BSc", "MSc", "PhD"]
education_enum = pl.Enum(education_level)

df = pl.DataFrame({
    "Name": ["Watson", "Justin", "Cheelo", "Joshua"],
    "Degree": pl.Series(["High school", "MSc", "PhD", "BSc"]),
})
df

shape: (4, 2)

Name	Degree
str	str
"Watson"	"High school"
"Justin"	"MSc"
"Cheelo"	"PhD"
"Joshua"	"BSc"

Filter with inequality

Because the Degree column is of type Enum, you can use inequality symbols to filter text values as if they were numbers. For example, you can filter for students who have a master’s degree (MSc) or higher like this:

(df
 .filter(pl.col("Degree") >= "MSc")
 )

shape: (2, 2)

Name	Degree
str	str
"Justin"	"MSc"
"Cheelo"	"PhD"

Check out my Polars course to learn more!