Pivot data without any aggregation in polars

unstack

You have a table with four rows and want to turn it into a table with two rows. In other words, you want to transform a long-format table into a wide-format table, which means you’ll have more columns. Below is a dataframe showing company departments and their budgets.

shape: (4, 2)

Department	Budget
str	i64
"Hospitality"	100
"Legal"	80
"Finance"	200
"Advertising"	120

Convert to wide format

To convert the dataframe above to a wide-format table without performing any aggregation, use the unstack method like this:

(df
 .unstack(step=2, how='horizontal')
 )

shape: (2, 4)

Department_0	Department_1	Budget_0	Budget_1
str	str	i64	i64
"Hospitality"	"Legal"	100	80
"Finance"	"Advertising"	200	120

The step parameter determines the number of rows you want to have in the final dataframe. You could also use pivot and leave the aggregate_function parameter blank to achieve the same result. However, unstack is often faster because it skips the grouping phase.

Learn more in my Polars course.