Indentation and Blocks

Purpose

Every block in Python — the body of a function, an if, a for, a class — is marked by indentation, not by a pair of braces or an end keyword. Lines at the same indentation level that follow a line ending in : belong to the same block.

Example

def describe(n: int) -> str:
    if n < 0:
        return "negative"
    return "non-negative"

Both return statements are part of describe’s body because they’re indented one level deeper than def. The return "non-negative" line is not part of the if block, because it sits back at the same indentation as if itself, not deeper.

Behavior notes

• A block is called a suite in the language grammar. A suite is either one or more indented lines following a :, or — for a short body — a single simple statement on the same line as the : (if n < 0: return "negative"). Lessons in this tier always use the indented form, since it reads more consistently as programs grow.
• Indentation must be consistent within a block, but Python does not mandate a specific width — four spaces is the overwhelming convention (and what every example on this site uses), codified for the standard library and most public code by PEP 8.
• Mixing tabs and spaces in the same file is an error in Python 3 — it raises TabError rather than silently guessing what you meant.
• An indented line where none is expected, or a missing indent where one is required, raises IndentationError, a subclass of SyntaxError — so, like any syntax error, it prevents the file from running at all rather than failing partway through.

Common mistakes

• Copy-pasting code from a source that used tabs into a file that uses spaces (or vice versa) — editors typically show both as the same visual width, so the mismatch is invisible until Python raises TabError.
• Assuming a line “looks” indented far enough because of a trailing comment or a wrapped editor window, rather than checking the actual leading whitespace.

Implementation notes

Indentation-defined blocks are part of the language grammar itself (see the specification link below), not a CPython-specific detail — every conforming Python implementation parses source the same way.