pandas-tibble

Display pandas DataFrames with R-style tibble formatting in the terminal.

What it does

• Tibble class - a DataFrame subclass that preserves all pandas functionality with automatic tibble-style display
• Prints DataFrames with column type annotations (<chr>, <int>, <dbl>, <date>, <datetime>)
• Shows dimensions at the top
• Distinguishes between date and datetime types
• Provides glimpse() for quick column overview
• All pandas operations work: .query(), .merge(), .groupby(), etc.

Installation

pip install pandas-tibble

For development:

git clone https://github.com/caspercrause/pandas-tibble.git
cd pandas-tibble
poetry install

Requirements

• Python >= 3.10.0
• pandas >= 2.0.0
• rich >= 13.0.0

Usage

Tibble class (Recommended)

Create a DataFrame subclass with automatic beautiful display:

from pandas_tibble import Tibble

# Create a Tibble - displays automatically!
tbl = Tibble({
    'product': ['A', 'B', 'A', 'B'],
    'revenue': [100, 150, 120, 180],
    'units': [10, 12, 11, 15]
})

# Just evaluate it - beautiful display
tbl

# All pandas operations work and return Tibbles!
tbl.query('revenue > 110')
tbl['profit'] = tbl['revenue'] * 0.2
tbl.groupby('product').sum()
tbl.merge(other_tbl, on='product')

Output:

# A DataFrame: 4 × 3

product  revenue  units
<chr>    <int>    <int>
A        100      10
B        150      12
A        120      11
B        180      15

tibble() function (for printing purposes only)

Display any DataFrame with type annotations:

import pandas as pd
from pandas_tibble import tibble
from datetime import date

df = pd.DataFrame({
    'start_date': [date(2024, 1, i) for i in range(1, 6)],
    'name': ['Alice', 'Bob', 'Charlie', 'Diana', 'Eve'],
    'value': [100, 200, 300, 400, 500]
})

tibble(df)

Output:

# A DataFrame: 5 × 3

start_date  name     value
<date>      <chr>    <int>
━━━━━━━━━━━━━━━━━━━━━━━━━━━
2024-01-01  Alice    100
2024-01-02  Bob      200
2024-01-03  Charlie  300
2024-01-04  Diana    400
2024-01-05  Eve      500

Options:

tibble(df, max_rows=20)         # Show more rows
tibble(df, show_index=True)     # Include index column

glimpse()

Quick column overview:

glimpse(df)

Output:

Rows: 5
Columns: 3

start_date  <date>  2024-01-01, 2024-01-02, 2024-01-03...
name        <chr>   Alice, Bob, Charlie...
value       <int>   100, 200, 300...

Type mapping

pandas dtype	Display	Format
object (string)	<chr>	as-is
int64	<int>	as-is
float64	<dbl>	2 decimals
bool	<lgl>	True/False
datetime64[ns]	<datetime>	YYYY-MM-DD HH:MM:SS
object (date)	<date>	YYYY-MM-DD
category	<fctr>	as-is
timedelta64[ns]	<timedelta>	as-is

Missing values display as NA.

Features

• ✅ Full pandas compatibility - all methods preserved
• ✅ Automatic beautiful display in terminal, Jupyter, VS Code
• ✅ Operations return Tibble objects (.query(), .merge(), .groupby(), etc.)
• ✅ Type annotations for all columns
• ✅ Works with existing DataFrames via Tibble(df) conversion

What it doesn't do

• Does not replicate R tibble's subsetting rules
• Does not support tibble construction syntax (use dict or DataFrame syntax)
• Does not modify how pandas operations work (100% compatible)

Similar Packages

Several packages provide enhanced DataFrame display or R-style functionality:

• rich-dataframe: Creates animated and pretty DataFrames using Rich, but is display-only and does not create a DataFrame subclass
• skimpy: Python port of R's skimr package, focuses on summary statistics rather than full tibble-style display
• tabulate: General table formatting library used by pandas' to_markdown(), provides formatting without DataFrame integration
• rich-tools: Helper functions to convert DataFrames to Rich tables for printing, not a DataFrame replacement
• datar: Full tidyverse port to Python with complete API replacement and heavy dependencies

What Sets pandas-tibble Apart

pandas-tibble is the only package that combines all of the following:

• DataFrame subclass that preserves all pandas methods and operations
• Automatic R-style tibble display without explicit print calls
• Lightweight implementation with only two dependencies (pandas and Rich)
• All operations return Tibble objects, maintaining consistency throughout your workflow
• Drop-in replacement for pandas DataFrames with zero API changes

Most packages are either display-only utilities or complete pandas replacements. pandas-tibble provides the best of both worlds: beautiful R-style display with full pandas compatibility.

License

MIT

Author

Casper Crause