funFrame is a small educational project that implements a DataFrame-style API using only Python's standard library.
This project was built for learning and experimentation. It is not optimized for performance and should not be used in production scenarios. The goal was to explore functional programming concepts and build a flexible, testable data manipulation library from scratch — without relying on external libraries like pandas or numpy.
- Pure Python: built using only the standard library (e.g.,
csv,array,copy,re) - A functional core API (pure functions for transformations)
- A convenient object-oriented wrapper:
DataFrameclass - Type-safe with support for
int,float,bool, andstrdata types - Read and parse CSV files
- Transform, filter, and combine columns
- Perform simple aggregations like
sum,product, andmean
git clone https://github.com/roman91DE/funFrame.gitThere are no external dependencies!
from funframe import DataFrame
# Create a DataFrame from a Python dict
df = DataFrame({
"a": [1, 2, 3],
"b": [4, 5, 6],
"c": ["x", "y", "z"]
})
# Add a new column
df2 = df.add_column("d", [7, 8, 9])
# Combine columns
df3 = df2.combine_columns("sum_ab", "a", "b", lambda x, y: x + y)
# Filter rows
filtered = df3.filter_on_column("a", lambda x: x > 1)
# Show data
filtered.show_all()from pathlib import Path
df = DataFrame.from_csv(Path("data.csv"))
df.show(5)Supports:
- Optional headers (
has_header=False) - Custom delimiters (
delimiter=";",delimiter="?", etc.) - Automatic type inference
The project comes with a suite of unit tests covering both functional and object-oriented APIs:
python test_funframe.pyThis project is intentionally simple and minimal:
- It's not fast (arrays are used, but no vectorization)
- It’s not comprehensive (no missing value support, no grouping, no joins)
- It's for learning and prototyping only
The implementation focuses on:
- Pure functions (where possible)
- Copy-on-write semantics (
deepcopy) - Transparent data structures (
dict[str, list|array]) - Explicit error handling and validation
- Want to understand how libraries like
pandaswork under the hood - Are learning Python and want a hands-on project
- Like building things from scratch instead of relying on magic ✨