pathlib cheat sheet: files and folders in Python

Older Python code glues paths together with strings and os.path.join. The pathlib module, part of the standard library, wraps a path in a Path object instead. You join parts with /, ask the object questions like .exists(), and read or write the file directly. The same code works on Windows, macOS and Linux because Path uses the right separator for the system.

from pathlib import Path

Building paths

Path.home()                     # your home folder
Path.cwd()                      # current working directory
Path(__file__).resolve().parent # folder of the running script

p = Path("reports") / "2026" / "summary.final.csv"
p.name        # 'summary.final.csv'
p.stem        # 'summary.final'
p.suffix      # '.csv'
p.suffixes    # ['.final', '.csv']
p.parent      # Path('reports/2026')
p.parts       # ('reports', '2026', 'summary.final.csv')

p.with_suffix(".json")      # reports/2026/summary.final.json
p.with_name("other.csv")    # reports/2026/other.csv
p.with_stem("summary.v2")   # reports/2026/summary.v2.csv

Tip: to open files that sit next to your script, build paths from Path(__file__).resolve().parent, not from the current directory. The current directory depends on where the user ran the command from.

Reading and writing

f = Path("output/logs/hello.txt")
f.parent.mkdir(parents=True, exist_ok=True)   # like mkdir -p

f.write_text("line one\nline two\n", encoding="utf-8")
text = f.read_text(encoding="utf-8")
lines = text.splitlines()

with f.open("a", encoding="utf-8") as fh:    # append, or stream big files
    fh.write("line three\n")

Path("data.bin").write_bytes(b"\x00\x01")
Path("data.bin").read_bytes()

Always pass encoding="utf-8" for text. Without it Python uses the platform default, which on some Windows setups is not UTF-8, and files with accents or emoji then fail to load.

Checking and inspecting

f.exists(), f.is_file(), f.is_dir()
f.stat().st_size          # size in bytes
f.stat().st_mtime         # last modified, as a timestamp
f.resolve()               # absolute path, symlinks resolved
f.relative_to("output")   # Path('logs/hello.txt')

Listing and searching

folder = Path("output")
for child in folder.iterdir():      # direct children only
    print(child)

sorted(Path(".").glob("*.py"))       # matching files in one folder
sorted(Path(".").rglob("*.txt"))     # recursive search

[p for p in folder.rglob("*") if p.is_file()]   # every file below

iterdir and glob return items in no particular order. Wrap them in sorted() when the order matters, for example when output must be reproducible.

Moving, copying and deleting

import shutil

new = f.rename(f.with_name("renamed.txt"))   # move or rename, returns new Path
shutil.copy2(new, new.with_name("copy.txt"))  # copy file with metadata
shutil.copytree("src_dir", "backup_dir")      # copy a whole folder

new.unlink()                       # delete a file
Path("maybe.txt").unlink(missing_ok=True)
Path("empty_dir").rmdir()          # only works on empty folders
shutil.rmtree("some_dir")          # deletes a folder and EVERYTHING in it

Deletions with unlink and shutil.rmtree are permanent, not a move to the trash. In automation scripts, print what would be deleted first (a “dry run”) and only delete when a flag such as --yes is given.

Converting

str(p)        # plain string, for libraries that want one
p.as_posix()  # always forward slashes, useful in URLs and configs
os.fspath(p)  # what most APIs call internally (import os first)

Most of the standard library and popular packages accept Path objects directly, so converting to str is rarely needed today.