Skip to content

Stable release · 0.2.0

Sorting that adapts to your data

BielSort is a stable sorting library for CPython. It accelerates favorable large integer lists with native Counting Sort or Radix Sort, then falls back to Python's Timsort whenever that is the safer or more compatible choice.

Install and start Explore the API Is BielSort a fit? View on PyPI View release

3.9–3.14CPython versions
5canonical functions
36prebuilt wheels
MITopen source license

Why BielSort?

Adaptive by design

The library inspects the workload and selects Counting Sort, Radix Sort, or Timsort instead of forcing one algorithm onto every input.

Python-compatible

Sorting remains stable. Version 0.2 can accelerate signed-int64 results from new-list key= calls; generic keys, in-place keys, non-integers, huge integers, and unsuitable ordered inputs retain Python's mature Timsort.

Native fast paths

The C extension targets exact Python integers in the signed 64-bit range and provides both new-list and in-place operations.

Start in seconds

python -m pip install bielsort

import bielsort

numbers = [8, -4, 10, 3, -4]
ordered = bielsort.sort(numbers)

print(ordered)  # [-4, -4, 3, 8, 10]
print(numbers)  # unchanged

import bielsort

numbers = [8, -4, 10, 3, -4]
bielsort.sort_in_place(numbers)

print(numbers)  # [-4, -4, 3, 8, 10]

Pick the operation that matches your code

Need Python BielSort
Create a new sorted list sorted(values) bielsort.sort(values)
Mutate an existing list values.sort() bielsort.sort_in_place(values)
See the selected strategy — bielsort.sort_with_strategy(values)
Mutate and see the strategy — bielsort.sort_in_place_with_strategy(values)
Inspect keyed algorithm and native memory — bielsort.sort_with_info(values, key=...)

A specialized tool, not a universal replacement

BielSort is most interesting for large list[int] workloads. Python's built-in sorting remains an excellent default for general objects, small inputs, nearly ordered data, in-place keys, and keyless reverse calls. See limits and compatibility before choosing it for a production workload.

Continue learning