You can use the **DataFrame.diff()** function to find the difference between two rows in a pandas DataFrame.

This function uses the following syntax:

**DataFrame.diff(periods=1, axis=0)**

where:

**periods:**The number of previous rows for calculating the difference.**axis:**Find difference over rows (0) or columns (1).

The following examples show how to use this function in practice.

**Example 1: Find Difference Between Each Previous Row**

Suppose we have the following pandas DataFrame:

import pandas as pd #create DataFrame df = pd.DataFrame({'period': [1, 2, 3, 4, 5, 6, 7, 8], 'sales': [12, 14, 15, 15, 18, 20, 19, 24], 'returns': [2, 2, 3, 3, 5, 4, 4, 6]}) #view DataFrame df period sales returns 0 1 12 2 1 2 14 2 2 3 15 3 3 4 15 3 4 5 18 5 5 6 20 4 6 7 19 4 7 8 24 6

The following code shows how to find the difference between every current row in a DataFrame and the previous row:

#add new column to represent sales differences between each row df['sales_diff'] = df['sales'].diff() #view DataFrame df period sales returns sales_diff 0 1 12 2 NaN 1 2 14 2 2.0 2 3 15 3 1.0 3 4 15 3 0.0 4 5 18 5 3.0 5 6 20 4 2.0 6 7 19 4 -1.0 7 8 24 6 5.0

Note that we can also find the difference between several rows prior. For example, the following code shows how to find the difference between each current row and the row that occurred three rows earlier:

#add new column to represent sales differences between current row and 3 rows earlier df['sales_diff'] = df['sales'].diff(periods=3) #view DataFrame df period sales returns sales_diff 0 1 12 2 NaN 1 2 14 2 NaN 2 3 15 3 NaN 3 4 15 3 3.0 4 5 18 5 4.0 5 6 20 4 5.0 6 7 19 4 4.0 7 8 24 6 6.0

**Example 2: Find Difference Based on Condition**

We can also filter the DataFrame to show rows where the difference between the current row and the previous row is less than or greater than some value.

For example, the following code returns only the rows where the value in the current row is **less than** the value in the previous row:

import pandas as pd #create DataFrame df = pd.DataFrame({'period': [1, 2, 3, 4, 5, 6, 7, 8], 'sales': [12, 14, 15, 13, 18, 20, 19, 24], 'returns': [2, 2, 3, 3, 5, 4, 4, 6]}) #find difference between each current row and the previous row df['sales_diff'] = df['sales'].diff() #filter for rows where difference is less than zero df = df[df['sales_diff']<0] #view DataFrame df period sales returns sales_diff 3 4 13 3 -2.0 6 7 19 4 -1.0

**Additional Resources**

How to Find Unique Values in Multiple Columns in Pandas

How to Filter a Pandas DataFrame by Column Values

How to Select Rows by Index in a Pandas DataFrame

Is it possible to compare sales of different period colum by column in the same data frame using Pandas, like sales for 2015 vs sales for 2014. Can’t get around the way it’s done on excel