Python:Pandas .groupby()

The Pandas DataFrame .groupby() function groups a DataFrame using a mapper or a series of columns and returns a GroupBy object. A range of methods, as well as custom functions, can be applied to GroupBy objects in order to combine or transform large amounts of data in these groups.

Pandas .groupby() Syntax

df.groupby(by=None, axis=0, level=None, as_index=True, sort=True, group_keys=True, observed=False, dropna=True)

Parameters:

• by: If a dictionary or Series is passed, the values will determine groups. If a list or ndarray with the same length as the selected axis is passed, the values will be used to form groups. A label or list of labels can be used to group by a particular column or columns.
• axis: Split along rows (0 or "index") or columns (1 or "columns").
• level: If the axis is a MultiIndex, group by a particular level or levels. Value is an integer or level name, or a sequence of them.
• as_index: Boolean value. True returns group labels as an index in aggregated output, and False returns labels as DataFrame columns.
• sort: Boolean value. True sorts the group keys.
• group_keys: Boolean value. If False, add group keys to index when calling apply.
• observed: Boolean value. If True, only show observed values for categorical groupers, otherwise show all values.
• dropna: Boolean value. If True, drop groups whose keys contain NA values. If False, NA will be used as a key for those groups.

Example 1: Group by Single Column Using .groupby()

This example uses .groupby() to group the data by a single column:

import pandas as pd

data = {
  'Region': ['East', 'West', 'East', 'South', 'West', 'South', 'East'],
  'Sales': [250, 200, 300, 400, 150, 500, 100]
}

df = pd.DataFrame(data)

result = df.groupby('Region')['Sales'].sum()

print(result)

Copy to clipboard
Copy to clipboard

Here is the output:

Region
East     650
South    900
West     350
Name: Sales, dtype: int64

Copy to clipboard
Copy to clipboard

Example 2: Group by Multiple Columns Using .groupby()

This example uses .groupby() to group the data by multiple columns:

import pandas as pd

data = {
  'Region': ['East', 'West', 'East', 'South', 'West', 'South', 'East'],
  'Product': ['A', 'B', 'A', 'B', 'A', 'A', 'B'],
  'Sales': [250, 200, 300, 400, 150, 500, 100]
}

df = pd.DataFrame(data)

result = df.groupby(['Region', 'Product'])['Sales'].sum()

print(result)

Copy to clipboard
Copy to clipboard

Here is the output:

Region  Product
East    A          550
        B          100
South   A          500
        B          400
West    A          150
        B          200
Name: Sales, dtype: int64

Copy to clipboard
Copy to clipboard

Codebyte Example: Using Aggregate Functions with Python’s .groupby()

This codebyte example uses .groupby() to group the data and then applies aggregate functions on the grouped data:

Visit us
Visit us
Hide code
Code
Output
Hide output
Loading...
Copy to your clipboard
Copy Codebyte
Run

Frequently Asked Questions

1. When should I use groupby in Pandas?

Use groupby when you want to split data into groups, apply a function, and combine results. Common operations include computing aggregates like sum, mean, or count per category.

2. Is Pandas groupby slow?

It can be slow for large datasets, especially if:

• You’re grouping by multiple columns.
• The dataset doesn’t fit in memory.
• You’re applying custom Python functions instead of built-ins.

For most medium-sized tasks, it’s fast enough. For massive data, look into more efficient libraries like Polars or Dask.

3. Is Polars groupby faster than Pandas?

Yes, often much faster. Polars is built in Rust and optimized for speed and parallelism. It can handle larger-than-memory data better and is ideal for performance-critical data tasks.

Example speed difference:

• Pandas: single-threaded.
• Polars: multi-threaded, faster groupby and aggregation.

If performance is a bottleneck, switching to Polars is worth considering.