How to reshape a Pandas DataFrame

How to reshape a Pandas DataFrame

To reshape a Pandas DataFrame, you can use the pivot() and pivot_table() functions.

These functions allow you to rearrange the data in a DataFrame, by creating new columns, rows, or levels based on the values in the original DataFrame.

Using pivot()

The pivot() method allows you to reshape a DataFrame by transforming the data into a new format, with different columns for each unique value in a specified column.

import pandas as pd

df = pd.DataFrame({'A': ['apple', 'banana', 'orange'],
                   'B': ['red', 'yellow', 'orange'],
                   'C': [1, 2, 3]})

# display the DataFrame

	      A     	B	    C
0	  apple	    red	    1
1	 banana	 yellow	    2
2	 orange	 orange	    3

This dataframe has three columns: A, B, and C. The pivot() method allows you to reshape the dataframe by transforming the data into a new format, with different columns for each unique value in a specified column.

For example, suppose we want to create a new DataFrame with columns A and B as the index, and C as the values.

We can use the pivot() method as follows:

# reshape the dataframe using the pivot() method
df_pivot = df.pivot(index='A', columns='B', values='C')

# display the resulting dataframe
df_pivot

This code will produce the following output:

B       red  yellow  orange
A
apple     1     NaN     NaN
banana  NaN       2     NaN
orange  NaN     NaN       3

As you can see, the pivot() method has transformed the data into a new format, with columns B as the index and C as the values.

The resulting dataframe has one column for each unique value in column B, and the values in column C are placed in the appropriate column based on the corresponding value in column B.

Using pivot_table()

For example, suppose we want to create a new dataframe with columns A and B as the index, and the sum of C as the values.

We can use the pivot_table() method as follows:

import pandas as pd

df = pd.DataFrame({'A': ['apple', 'banana', 'orange'],
                   'B': ['red', 'yellow', 'orange'],
                   'C': [1, 2, 3]})

# reshape the dataframe using the pivot_table() method
df_pivot_table = df.pivot_table(index='A', columns='B', values='C', aggfunc=sum)

# display the resulting dataframe
df_pivot_table

B       red  yellow  orange
A
apple     1     NaN     NaN
banana  NaN       2     NaN
orange  NaN     NaN       3

The pivot_table() method has aggregated the data using the sum() function, and has transformed the data into a new format with columns B as the index and the sum of C as the values.

The resulting dataframe has one column for each unique value in column B, and the sum of the values in column C are placed in the appropriate column based on the corresponding value in column B.