
Nobody remembers the exact arguments to csv.DictWriter, and that's fine. Most everyday Python is made of small moves like that: read a file, count things, fix a date, run a command.
Here are 20 of them as micro-recipes. Each one is short, runs as written, and comes with one note about the thing that usually trips people up. Everything uses the standard library, so there's nothing to install.
Versions: every snippet was run on Python 3.12. The recipes that need a recent version say so.
Files and data
1. Read and write a text file
from pathlib import Path
path = Path("notes.txt")
path.write_text("hello\nworld\n", encoding="utf-8")
text = path.read_text(encoding="utf-8")
lines = text.splitlines()
print(lines) # ['hello', 'world']Heads up: always pass
encoding. The default depends on the operating system, which is how scripts that work on your laptop break on a colleague's Windows machine. For huge files, loop overpath.open()instead of reading everything at once.
2. Find files recursively
from pathlib import Path
for path in sorted(Path("src").rglob("*.py")):
print(path.stem, path.suffix, path.stat().st_size)rglob searches every subfolder. Use glob("*.py") to stay in one folder. Each Path also gives you .name, .stem, .suffix and .parent for free, so you never need to slice strings.
3. Read and write JSON
import json
from pathlib import Path
config = {"name": "demo", "retries": 3}
Path("config.json").write_text(json.dumps(config, indent=2), encoding="utf-8")
loaded = json.loads(Path("config.json").read_text(encoding="utf-8"))
print(loaded["retries"]) # 3Heads up:
dumpsandloadswork with strings;dumpandloadwork with open files. If your data contains dates, passdefault=strtodumps, or it raises aTypeError.
4. Read and write CSV as dictionaries
import csv
rows = [{"name": "Asha", "email": "[email protected]"}]
with open("users.csv", "w", newline="", encoding="utf-8") as f:
writer = csv.DictWriter(f, fieldnames=["name", "email"])
writer.writeheader()
writer.writerows(rows)
with open("users.csv", newline="", encoding="utf-8") as f:
for row in csv.DictReader(f):
print(row["name"], row["email"]) # Asha [email protected]Heads up:
newline=""is required, or you get blank lines on Windows. Every value comes back as a string, so convert numbers yourself withint(row["age"]).
Scripts and the command line
5. Add command-line arguments
import argparse
parser = argparse.ArgumentParser(description="Resize images")
parser.add_argument("folder")
parser.add_argument("--width", type=int, default=800)
parser.add_argument("--dry-run", action="store_true")
args = parser.parse_args()
print(args.folder, args.width, args.dry_run)Run it as python resize.py photos --width 400 --dry-run. You get type conversion, defaults and a free --help page. A dash in the flag becomes an underscore in the code: --dry-run is args.dry_run.
6. Run a shell command and capture the output
import subprocess
result = subprocess.run(
["git", "status", "--short"],
capture_output=True,
text=True,
check=True,
timeout=30,
)
print(result.stdout)Heads up: pass the command as a list, not one long string. Avoid
shell=Truewhenever any part of the command comes from user input.check=TrueraisesCalledProcessErroron a non-zero exit code, so failures don't pass silently.
Lists and dicts
7. Count things
from collections import Counter
words = "the cat saw the dog and the cat ran".split()
counts = Counter(words)
print(counts.most_common(2)) # [('the', 3), ('cat', 2)]A Counter behaves like a dict, and missing keys return 0 instead of raising an error.
8. Group things by a key
from collections import defaultdict
from pathlib import Path
by_ext = defaultdict(list)
for name in ["a.py", "b.txt", "c.py"]:
by_ext[Path(name).suffix].append(name)
print(dict(by_ext)) # {'.py': ['a.py', 'c.py'], '.txt': ['b.txt']}No more if key not in d: d[key] = []. Use defaultdict(int) when you want running totals instead of lists.
9. Merge dictionaries
defaults = {"timeout": 10, "retries": 3}
overrides = {"timeout": 30}
settings = defaults | overrides
print(settings) # {'timeout': 30, 'retries': 3}The right side wins on conflicts. Needs Python 3.9+. The merge is shallow, so nested dicts are replaced, not combined.
10. Remove duplicates but keep the order
items = ["b", "a", "b", "c", "a"]
unique = list(dict.fromkeys(items))
print(unique) # ['b', 'a', 'c']set(items) also removes duplicates, but it scrambles the order. Dicts keep insertion order, so this trick doesn't. Items must be hashable.
11. Flatten a list of lists
from itertools import chain
nested = [[1, 2], [3], [4, 5]]
flat = list(chain.from_iterable(nested))
print(flat) # [1, 2, 3, 4, 5]This flattens exactly one level, which is almost always what you want.
12. Split a list into chunks
from itertools import batched
for batch in batched(range(7), 3):
print(batch)
# (0, 1, 2)
# (3, 4, 5)
# (6,)Perfect for sending records to an API in groups of 100. Needs Python 3.12+, and each batch is a tuple. The last one can be shorter.
13. Sort by more than one field
people = [("Ben", 31), ("Asha", 25), ("Cara", 31)]
ranked = sorted(people, key=lambda p: (-p[1], p[0]))
print(ranked) # [('Ben', 31), ('Cara', 31), ('Asha', 25)]The key returns a tuple: oldest first (the minus sign flips the number), then name A to Z. This only works for numbers. For descending text, sort twice, since Python's sort is stable.
14. Pair two lists without silent data loss
names = ["Asha", "Ben", "Cara"]
scores = [90, 85]
for name, score in zip(names, scores, strict=True):
print(name, score)
# Asha 90
# Ben 85
# ValueError: zip() argument 2 is shorter than argument 1Plain zip quietly stops at the shortest list, so Cara would vanish without a warning. strict=True turns that into an error. Needs Python 3.10+.
Text
15. Pull fields out of a string with regex
import re
line = "2026-11-22 ERROR payment failed order=1042"
pattern = r"(?P<date>\d{4}-\d{2}-\d{2}) (?P<level>[A-Z]+) .*order=(?P<order>\d+)"
match = re.search(pattern, line)
if match:
print(match["date"], match["level"], match["order"]) # 2026-11-22 ERROR 1042Heads up: use raw strings (
r"...") for patterns, and always check forNonebefore using the match. Named groups make the code readable six months later. Usere.findallwhen you want every match, not just the first.
16. Format numbers with f-strings
price = 1234567.891
x = 7
print(f"{price:,.2f}") # 1,234,567.89
print(f"{0.256:.1%}") # 25.6%
print(f"{42:05d}") # 00042
print(f"{'name':<10}|") # name |
print(f"{x=}") # x=7The part after the colon is a format spec: , adds thousands separators, .2f is two decimals, % multiplies by 100, <10 pads to ten characters. The x= form prints the name and the value, which is great for quick debugging.
Dates and time
17. Parse, format and shift dates
from datetime import datetime, timedelta
dt = datetime.strptime("2026-11-22 14:30", "%Y-%m-%d %H:%M")
print(dt.strftime("%d %b %Y")) # 22 Nov 2026
print(dt + timedelta(days=30)) # 2026-12-22 14:30:00
other = datetime.fromisoformat("2026-12-25T09:00:00")
print((other - dt).days) # 32strptime turns text into a date, strftime turns a date into text. For ISO-style strings, fromisoformat needs no format code at all. Subtracting two datetimes gives a timedelta.
18. Convert between time zones
from datetime import datetime
from zoneinfo import ZoneInfo
meeting = datetime(2026, 11, 22, 9, 0, tzinfo=ZoneInfo("America/New_York"))
print(meeting.astimezone(ZoneInfo("Asia/Kolkata"))) # 2026-11-22 19:30:00+05:30Heads up: store and compare times in UTC, and convert only for display. Avoid
datetime.utcnow(). It's deprecated since Python 3.12 and returns a naive datetime. Usedatetime.now(ZoneInfo("UTC"))instead. On Windows you may needpip install tzdata.
Speed
19. Run slow I/O tasks in parallel
from concurrent.futures import ThreadPoolExecutor
from urllib.request import urlopen
def fetch_status(url):
with urlopen(url, timeout=10) as response:
return response.status
urls = ["https://example.com", "https://python.org"]
with ThreadPoolExecutor(max_workers=8) as pool:
statuses = list(pool.map(fetch_status, urls))
print(statuses) # [200, 200]Threads shine when your code waits on the network or disk. Results come back in the same order as the input. If a call fails, the exception is raised when you read that result. For heavy number crunching, switch to ProcessPoolExecutor.
20. Cache a function's results
from functools import cache
@cache
def fib(n):
return n if n < 2 else fib(n - 1) + fib(n - 2)
print(fib(80)) # 23416728348467685Add @cache to any function that is slow and gets called repeatedly with the same arguments. Arguments must be hashable. The cache grows forever, so use lru_cache(maxsize=1000) for long-running programs, and call fib.cache_clear() to reset it.
The cheat sheet
I need to... | Reach for |
|---|---|
Read or write a file |
|
Find files |
|
Handle JSON |
|
Handle CSV |
|
Build a CLI |
|
Run a command |
|
Count things |
|
Group things |
|
Merge dicts | the dict union operator |
Dedupe in order |
|
Flatten a list |
|
Chunk a list |
|
Sort by several fields |
|
Pair lists safely |
|
Extract from text |
|
Format numbers | f-string format specs |
Parse dates |
|
Convert time zones |
|
Parallelize I/O |
|
Cache results |
|
Pick five you use most, paste them into a snippets file, and stop memorizing. For everything else, help() and the official docs are one keystroke away.
Comments(0)
Join the discussion by logging into your account.
No comments yet. Be the first to comment!