The reading room

Python explainers

Explore explanations and worked examples. Search by topic, or follow a track to combine reading with practice.

Prefer a guided path?

Explore Python tracks

177 explainers Showing 25–48

Catch Several Exceptions Without Parentheses

Python 3.14 lets you list several exception types in an except clause without parentheses. Learn when you still need them, and which old spelling now means something else.

About 3 min Read explainer
Flatten and Merge with Unpacking in Comprehensions

Python 3.15 lets you write * and ** inside a comprehension to flatten lists and merge dictionaries. See where the star goes, the trap with function calls, and what to write on older versions.

About 4 min Read explainer
Speed Up Startup with Lazy Imports

Python 3.15's lazy import loads a module the first time your code uses it. Learn where you can write it, what moves from startup to first use, and how to control it for a whole program.

About 4 min Read explainer
Bring Old Python Code Up to Date

Look up the newer spelling of common old Python code and what to check before each change. Learn to update old code without changing what it does.

About 4 min Read explainer
Write Code That Runs on Several Python Versions

New syntax, new library names and changed behavior each break code on older Python versions in a different way. Learn the fix for each, and how to test every version you promise to support.

About 4 min Read explainer
Count Repeated Values with Counter

Count words or stock with Counter, rank ties in the order you choose, and keep negative changes when a report needs them.

About 4 min Read explainer
Handle Missing Keys with get, setdefault and defaultdict

Group values under a key with setdefault or defaultdict, learn which lookups add a key, and hand callers a plain dict.

About 4 min Read explainer
Queues and Sliding Windows with deque

Use a deque as a fast queue, keep only the most recent items with maxlen, and see what a rolling median still costs.

About 4 min Read explainer
Priority Queues, Sorted Lists and Summary Statistics

Build a priority queue with heapq that keeps arrival order for ties, find positions in a sorted list with bisect, and pick the right summary from statistics.

About 4 min Read explainer
Layer Configuration with ChainMap

Look up a setting across default, environment and command-line layers without copying them, and know which layer a write changes.

About 4 min Read explainer
When to Use OrderedDict Instead of dict

A plain dict keeps insertion order. Use OrderedDict when you move keys around, remove the oldest entry, or need order to count in equality.

About 4 min Read explainer
Generators and One-Shot Iterators

Write a generator with yield, see that its body runs only when you ask for a value, and avoid looping twice over an iterator that is already used up.

About 4 min Read explainer
Take the First Few Items with chain and islice

Join several sources into one stream with chain, stop after n items with islice, and read no more input than you need.

About 3 min Read explainer
Group Adjacent Items with groupby

Group adjacent items that share a key with groupby. Sort first when you want one group per key, and use each group before you move to the next.

About 3 min Read explainer
Split a Stream into Batches with batched

Send records in fixed-size batches with itertools.batched, keep the short last batch, and reject a bad batch size at call time.

About 3 min Read explainer
Adjacent Pairs and Running Totals with pairwise and accumulate

Compute day-to-day changes with pairwise and running totals or peaks with accumulate, and know when looping over a source twice is safe.

About 4 min Read explainer
Split an Iterator with tee and Combine Inputs with product

Give two loops their own view of one iterator with tee, and count the tuples product will make before you loop over them.

About 4 min Read explainer
Repeat, Pad and Filter Streams with itertools

Use cycle, repeat, zip_longest, takewhile, dropwhile, filterfalse, starmap and compress, and see which items each one reads, keeps or drops.

About 4 min Read explainer
Reuse Earlier Results with cache and lru_cache

Cache a function's results, keep a typed public function in front of the cache, and decide how long the results should live.

About 4 min Read explainer
Cache Method Results Without Keeping Objects Alive

See why lru_cache on a method keeps objects in memory, and give each object its own cache or cache a plain function instead.

About 4 min Read explainer
Compute an Attribute Once with cached_property

Compute an attribute on its first read, clear it when its inputs change, and cache by hand in a class that uses slots.

About 3 min Read explainer
Choose a Function by Argument Type with singledispatch

Replace a chain of isinstance checks with handlers registered per type, so other modules can add types without editing your function.

About 3 min Read explainer
Write Typed Decorators with wraps and ParamSpec

Learn what a decorator does, keep the wrapped function's name with wraps, and keep its parameter and return types with TypeVar and ParamSpec.

About 4 min Read explainer
Preset Arguments with partial and partialmethod

Make a new callable with some arguments filled in, put presets on a class with partialmethod and staticmethod, and type a retry wrapper.

About 4 min Read explainer

Catch Several Exceptions Without Parentheses

Flatten and Merge with Unpacking in Comprehensions

Speed Up Startup with Lazy Imports

Bring Old Python Code Up to Date

Write Code That Runs on Several Python Versions

Count Repeated Values with Counter

Handle Missing Keys with get, setdefault and defaultdict

Queues and Sliding Windows with deque

Priority Queues, Sorted Lists and Summary Statistics

Layer Configuration with ChainMap

When to Use OrderedDict Instead of dict

Generators and One-Shot Iterators

Take the First Few Items with chain and islice

Group Adjacent Items with groupby

Split a Stream into Batches with batched

Adjacent Pairs and Running Totals with pairwise and accumulate

Split an Iterator with tee and Combine Inputs with product

Repeat, Pad and Filter Streams with itertools

Reuse Earlier Results with cache and lru_cache

Cache Method Results Without Keeping Objects Alive

Compute an Attribute Once with cached_property

Choose a Function by Argument Type with singledispatch

Write Typed Decorators with wraps and ParamSpec

Preset Arguments with partial and partialmethod