Convert your data from wide to long format in polars

unpivot

Wide-format data (data with many columns) is good for presentation, but for computation, long format is best. That’s what computers love! Below is a dataframe showing software sold on different days of the week.

shape: (4, 6)

Item	Mon	Tue	Wed	Thu	Fri
str	i64	i64	i64	i64	i64
"Photoshop"	25	30	28	26	27
"CorelDRAW"	18	20	22	21	19
"Netscape"	22	24	19	23	25
"Encarta"	15	17	20	18	16

Convert to long format

To convert the data to long format, you can use the Polars method unpivot, which is the opposite of pivot, like this:

(df
 .unpivot(index='Item',
          variable_name='Day',
          value_name='Units')
 )

shape: (20, 3)

Item	Day	Units
str	str	i64
"Photoshop"	"Mon"	25
"CorelDRAW"	"Mon"	18
"Netscape"	"Mon"	22
"Encarta"	"Mon"	15
"Photoshop"	"Tue"	30
…	…	…
"Encarta"	"Thu"	18
"Photoshop"	"Fri"	27
"CorelDRAW"	"Fri"	19
"Netscape"	"Fri"	25
"Encarta"	"Fri"	16

Now the dataframe has fewer columns but many more rows, making it long format.

Join other students learning in my Polars course.