zipp

3.23.0 · active · verified Sat Mar 28

zipp is the official backport of the pathlib-compatible zipfile.Path object from the Python standard library. It provides a Traversable/pathlib-like interface for navigating and reading ZIP archives (zipp.Path), with new features introduced here first and later merged into CPython. Current version is 3.23.0, released June 2025, with a very active cadence of roughly one release per month maintained by Jason R. Coombs (jaraco).

Warnings

breaking zipp.Path.open() changed to text mode by default and dropped positional pwd and force_zip64 parameters.
Fix: Pass mode='rb' explicitly for binary reads. Remove any positional pwd argument; keyword args still pass through to io.TextIOWrapper.
breaking Path.joinpath() dropped the 'add' keyword argument; it now only accepts arbitrary positional path segments.
Fix: Replace joinpath(add='sub/path') with joinpath('sub/path') or the / operator (root / 'sub' / 'path').
breaking Requires Python >=3.9 as of recent releases. Older zipp releases supported Python 3.6/3.7/3.8 but those branches are no longer maintained.
Fix: Pin to zipp<3.20 if you must support Python 3.8. For Python 3.7, pin to zipp<3.5.
gotcha Directory entries from iterdir() always end with a trailing slash in their string representation. Comparing path names with string equality will fail unless you strip the slash.
Fix: Use path.name (strips the slash) rather than str(path) when comparing directory names.
gotcha Passing a ZipFile object to zipp.Path mutates the ZipFile in place (converts it to a CompleteDirs subclass). The caller's reference is affected. This is by design to prevent double-close bugs.
Fix: Do not rely on the original ZipFile object type after wrapping it in zipp.Path; use the zipp.Path interface exclusively for all navigation.
gotcha Path.glob('**') previously did not match directories; since 3.20 it does. Code that relied on glob returning only files may now receive unexpected directory entries.
Fix: Filter results with item.is_file() when you want only files: [p for p in root.glob('**/*.txt') if p.is_file()].
gotcha zipp is commonly pulled in as a transitive dependency (e.g. by importlib-metadata, pip, pytest). Pinning it too tightly in lock files can block upgrades of those tools.
Fix: Declare a loose lower bound (zipp>=3.9) rather than an exact pin unless you have a specific compatibility reason.

Install

pip install zipp Latest stable

Imports

Path
```
import zipp; path = zipp.Path('archive.zip')
```
Use zipp.Path, not zipfile.Path — zipp always carries the latest features ahead of the stdlib backport.
zipp.compat.overlay
```
import zipp.compat.overlay
```
The zipfile compatibility overlay submodule; available since zipp 3.19. Do not attempt to import it on older versions.

Quickstart

Create an in-memory ZIP, wrap it with zipp.Path, iterate entries, read file content, and glob for files — all using a pathlib-style API.

import io
import zipfile
import zipp

# Build an in-memory zip
data = io.BytesIO()
with zipfile.ZipFile(data, 'w') as zf:
    zf.writestr('hello.txt', 'Hello, zipp!')
    zf.writestr('sub/world.txt', 'World!')
    zf.filename = 'example.zip'

# Navigate with zipp.Path (pathlib-style)
root = zipp.Path(data)
for item in root.iterdir():
    print(item.name, item.is_file())

# Read a nested file
file_path = root / 'hello.txt'
print(file_path.read_text())  # 'Hello, zipp!'

# Glob support
for txt in root.glob('**/*.txt'):
    print(txt)

view raw JSON →