python-peps/pep-0622.rst

PEP: 622
Title: Structural Pattern Matching
Version: $Revision$
Last-Modified: $Date$
Author: Brandt Bucher <brandtbucher@gmail.com>,
        Daniel F Moisset <dfmoisset@gmail.com>,
        Tobias Kohn <kohnt@tobiaskohn.ch>,
        Ivan Levkivskyi <levkivskyi@gmail.com>,
        Guido van Rossum <guido@python.org>,
        Talin <viridia@gmail.com>
BDFL-Delegate:
Discussions-To: Python-Dev <python-dev@python.org>
Status: Draft
Type: Standards Track
Content-Type: text/x-rst
Created: 23-Jun-2020
Python-Version: 3.10
Post-History: 23-Jun-2020, 8-Jul-2020
Resolution:


Abstract
========

This PEP proposes to add a pattern matching statement to Python,
inspired by similar syntax found in Scala and many other languages.

The pattern syntax builds on Python’s existing syntax for sequence
unpacking (e.g., ``a, b = value``), but is wrapped in a ``match``
statement which compares its subject to several different “shapes”
until one is found that fits. In addition to specifying the shape of a
sequence to be unpacked, patterns can also specify the shape to be a
mapping with specific keys, an instance of a given class with (optionally) specific
attributes, a specific value, or a wildcard. Patterns can be composed
in several ways.

Syntactically, a ``match`` statement contains a *subject* expression
and one or more ``case`` clauses, where each case clause specifies a
pattern (the overall shape to be matched), an optional “guard” (a
condition to be checked if the pattern matches), and a code block to
be executed if the case clause is selected.

The rest of the PEP motivates why we believe pattern matching makes a
good addition to Python, explains our design choices, and contains a
precise syntactic and runtime specification. We also give guidance for
static type checkers (and one small addition to the ``typing`` module)
and discuss the main objections and alternatives that have been
brought up during extensive discussion of the proposal, both within
the group of authors and in the python-dev community. Finally, we
discuss some possible extensions that might be considered in the
future, once the community has ample experience with the currently
proposed syntax and semantics.


Overview
========

Since patterns are a new syntactical category with their own rules
and special cases, and since they mix input (given values) and output
(captured variables) in novel ways, they require a bit of getting used
to. It is the experience of the authors that this happens quickly when
a brief introduction to the basic concepts such as the following is
presented. Note that this section is not intended to be complete or
perfectly accurate.

A new syntactic construct called *pattern* is
introduced. Syntactically, patterns look like a subset of expressions;
the following are patterns:

- ``[first, second, *rest]``
- ``Point2d(x, 0)``
- ``{"name": "Bruce", "age": age}``
- ``42``

The above look like examples of object construction. A constructor
takes some values as parameters and builds an object from those
components. But as a pattern the above mean the inverse operation of
construction, which we call *destructuring*: it takes a subject value
and extracts its components. The syntactic similarity between
construction and destructuring is intentional and follows the existing
Pythonic style which makes assignment targets (write contexts) look
like expressions (read contexts). Pattern matching never creates
objects, in the same way that ``[a, b] = my_list`` doesn't create a
new ``[a, b]`` list, nor reads the values of ``a`` and ``b``.

The intuition we are trying to build in users as they learn this is
that matching a pattern to a subject binds the free variables (if any)
to subject components in a way that reflects the original
subject when read as an expression. During this process,
the structure of the pattern may not fit the subject, in which case
the matching *fails*. For example, matching the pattern ``Point2d(x,
0)`` to the subject ``Point2d(3, 0)`` successfully matches and binds
``x`` to ``3``. However, if the subject is ``[3, 0]`` the match fails
because a ``list`` is not a ``Point2d``. And if the subject is
``Point2D(3, 3)`` the match fails because its second coordinate is not
``0``.

The ``match`` statement tries to match each of the
patterns in its ``case`` clauses with a single subject. At the first
successful match, the variables in the pattern are assigned and a
corresponding block is executed. Each of the multiple branches of this
conditional statement can also have a boolean condition as a *guard*.

Here's an example of a match statement, used to define a function
building 3D points that can accept as input either tuples of size 2 or
3, or existing (2D or 3D) points::

    def make_point_3d(pt):
        match pt:
            case (x, y):
                return Point3d(x, y, 0)
            case (x, y, z):
                return Point3d(x, y, z)
            case Point2d(x, y):
                return Point3d(x, y, 0)
            case Point3d(_, _, _):
                return pt
            case _:
                raise TypeError("not a point we support")

Writing this function in the traditional fashion would require several
``isinstance()`` checks, one or two ``len()`` calls, and a more
convoluted control flow. While the ``match`` version translates into
similar code under the hood, to a reader familiar with patterns it is
much clearer.


Rationale and Goals
===================

Python programs frequently need to handle data which varies in type,
presence of attributes/keys, or number of elements. Typical examples
are operating on nodes of a mixed structure like an AST, handling UI
events of different types, processing structured input (like
structured files or network messages), or “parsing” arguments for a
function that can accept different combinations of types and numbers
of parameters. In fact, the classic 'visitor' pattern is an example of this,
done in an OOP style -- but matching makes it much less tedious to write.

Much of the code to do so tends to consist of complex chains of nested
``if``/``elif`` statements, including multiple calls to ``len()``,
``isinstance()`` and index/key/attribute access. Inside those branches
users sometimes need to destructure the data further to extract the
required component values, which may be nested several objects deep.

Pattern matching as present in many other languages provides an
elegant solution to this problem. These range from statically compiled
functional languages like F# and Haskell, via mixed-paradigm languages
like Scala [4]_ and Rust [3]_, to dynamic languages like Elixir and
Ruby, and is under consideration for JavaScript. We are indebted to
these languages for guiding the way to Pythonic pattern matching, as
Python is indebted to so many other languages for many of its
features: many basic syntactic features were inherited from C,
exceptions from Modula-3, classes were inspired by C++, slicing came
from Icon, regular expressions from Perl, decorators resemble Java
annotations, and so on.

The usual logic for operating on heterogeneous data can be summarized
in the following way:

- Some analysis is done on the *shape* (type and components) of the
  data: This could involve ``isinstance()`` or ``len()`` calls and/or extracting
  components (via indexing or attribute access) which are checked for
  specific values or conditions.
- If the shape is as expected, some more components are possibly
  extracted and some operation is done using the extracted values.

Take for example `this piece of the Django web framework
<https://github.com/django/django/blob/5166097d7c80cab757e44f2d02f3d148fbbc2ff6/django/db/models/enums.py#L13>`_::

    if (
        isinstance(value, (list, tuple)) and
        len(value) > 1 and
        isinstance(value[-1], (Promise, str))
    ):
        *value, label = value
        value = tuple(value)
    else:
        label = key.replace('_', ' ').title()

We can see the shape analysis of the ``value`` at the top, following
by the destructuring inside.

Note that shape analysis here involves checking the types both of the
container and of one of its components, and some checks on its number
of elements. Once we match the shape, we need to decompose the
sequence. With the proposal in this PEP, we could rewrite that code
into this::

    match value:
        case [*v, label := (Promise() | str())] if v:
            value = tuple(v)
        case _:
            label = key.replace('_', ' ').title()

This syntax makes much more explicit which formats are possible for
the input data, and which components are extracted from where. You can
see a pattern similar to list unpacking, but also type checking: the
``Promise()`` pattern is not an object construction, but represents
anything that's an instance of ``Promise``. The pattern operator ``|``
separates alternative patterns (not unlike regular expressions or EBNF
grammars), and ``_`` is a wildcard. (Note that the match syntax used
here will accept user-defined sequences, as well as lists and tuples.)

In some occasions, extraction of information is not as relevant as
identifying structure. Take the following example from the
`Python standard library
<https://github.com/python/cpython/blob/c4cacc8/Lib/lib2to3/fixer_util.py#L158>`_::

    def is_tuple(node):
        if isinstance(node, Node) and node.children == [LParen(), RParen()]:
            return True
        return (isinstance(node, Node)
                and len(node.children) == 3
                and isinstance(node.children[0], Leaf)
                and isinstance(node.children[1], Node)
                and isinstance(node.children[2], Leaf)
                and node.children[0].value == "("
                and node.children[2].value == ")")

This example shows an example of finding out the "shape" of the data
without doing significant extraction. This code is not very easy to
read, and the intended shape that this is trying to match is not
evident. Compare with the updated code using the proposed syntax::

    def is_tuple(node: Node) -> bool:
        match node:
            case Node(children=[LParen(), RParen()]):
                return True
            case Node(children=[Leaf(value="("), Node(), Leaf(value=")")]):
                return True
            case _:
                return False

Note that the proposed code will work without any modifications to the
definition of ``Node`` and other classes here. As shown in the
examples above, the proposal supports not just unpacking sequences, but
also doing ``isinstance`` checks (like ``LParen()`` or ``str()``),
looking into object attributes (``Leaf(value="(")`` for example) and
comparisons with literals.

That last feature helps with some kinds of code which look more like
the "switch" statement as present in other languages::

    match response.status:
        case 200:
            do_something(response.data)  # OK
        case 301 | 302:
            retry(response.location)  # Redirect
        case 401:
            retry(auth=get_credentials())  # Login first
        case 426:
            sleep(DELAY)  # Server is swamped, try after a bit
            retry()
        case _:
            raise RequestError("we couldn't get the data")

Although this will work, it's not necessarily what the proposal is
focused on, and the new syntax has been designed to best support the
destructuring scenarios.

See the `syntax`_ sections below for a more detailed specification.

We propose that destructuring objects can be customized by a new
special ``__match_args__`` attribute. As part of this PEP we specify
the general API and its implementation for some standard library
classes (including named tuples and dataclasses). See the `runtime`_
section below.

Finally, we aim to provide comprehensive support for static type
checkers and similar tools. For this purpose, we propose to introduce
a ``@typing.sealed`` class decorator that will be a no-op at runtime
but will indicate to static tools that all sub-classes of this class
must be defined in the same module. This will allow effective static
exhaustiveness checks, and together with dataclasses, will provide
basic support for algebraic data types [2]_. See the `static checkers`_
section for more details.


.. _syntax:

Syntax and Semantics
====================

Patterns
--------

The **pattern** is a new syntactical construct, that could be considered a loose
generalization of assignment targets. The key properties of a pattern are what
types and shapes of subjects it accepts, what variables it captures and how
it extracts them from the subject. For example the pattern ``[a, b]`` matches
only sequences of exactly 2 elements, extracting the first element into ``a``
and the second one into ``b``.

This PEP defines several types of patterns. These are certainly not the
only possible ones, so the design decision was made to choose a subset of
functionality that is useful now but conservative. More patterns can be added
later as this feature gets more widespread use. See the `rejected ideas`_
and `deferred ideas`_ sections for more details.

The patterns listed here are described in more detail below, but summarized
together in this section for simplicity:

- A **literal pattern** is useful to filter constant values in a structure.
  It looks like a Python literal (including some values like ``True``,
  ``False`` and ``None``). It only matches objects equal to the literal, and
  never binds.
- A **capture pattern** looks like ``x`` and is equivalent to an identical
  assignment target: it always matches and binds the variable
  with the given (simple) name.
- The **wildcard pattern** is a single underscore: ``_``. It always matches,
  but does not capture any variable (which prevents interference with other
  uses for ``_`` and allows for some optimizations).
- A **constant value pattern** works like the literal but for certain named
  constants. Note that it must be a qualified (dotted) name, given the possible
  ambiguity with a capture pattern. It looks like ``Color.RED`` and
  only matches values equal to the corresponding value. It never binds.
- A **sequence pattern** looks like ``[a, *rest, b]`` and is similar to
  a list unpacking. An important difference is that the elements nested
  within it can be any kind of patterns, not just names or sequences.
  It matches only sequences of appropriate length, as long as all the sub-patterns
  also match. It makes all the bindings of its sub-patterns.
- A **mapping pattern** looks like ``{"user": u, "emails": [*es]}``. It matches
  mappings with at least the set of provided keys, and if all the
  sub-patterns match their corresponding values. It binds whatever the
  sub-patterns bind while matching with the values corresponding to the keys.
  Adding ``**rest`` at the end of the pattern to capture extra items is allowed.
- A **class pattern** is similar to the above but matches attributes instead
  of keys. It looks like ``datetime.date(year=y, day=d)``. It matches
  instances of the given type, having at least the specified
  attributes, as long as the attributes match with the corresponding
  sub-patterns. It binds whatever the sub-patterns bind when matching with the
  values of
  the given attributes. An optional protocol also allows matching positional
  arguments.
- An **OR pattern** looks like ``[*x] | {"elems": [*x]}``. It matches if any
  of its sub-patterns match. It uses the binding for the leftmost pattern
  that matched.
- A **walrus pattern** looks like ``d := datetime(year=2020, month=m)``. It
  matches only
  if its sub-pattern also matches. It binds whatever the sub-pattern match does, and
  also binds the named variable to the entire object.

The ``match`` statement
-----------------------

A simplified, approximate grammar for the proposed syntax is::

  ...
  compound_statement:
      | if_stmt
      ...
      | match_stmt
  match_stmt: "match" expression ':' NEWLINE INDENT case_block+ DEDENT
  case_block: "case" pattern [guard] ':' block
  guard: 'if' expression
  pattern: walrus_pattern | or_pattern
  walrus_pattern: NAME ':=' or_pattern
  or_pattern: closed_pattern ('|' closed_pattern)*
  closed_pattern:
      | literal_pattern
      | capture_pattern
      | wildcard_pattern
      | constant_pattern
      | sequence_pattern
      | mapping_pattern
      | class_pattern

See `Appendix A`_ for the full, unabridged grammar. The simplified grammars in
this section are there for helping the reader, not as a full specification.

We propose that the match operation should be a statement, not an expression.
Although in

many languages it is an expression, being a statement better suits the general
logic of Python syntax. See `rejected ideas`_ for more discussion.
The allowed patterns are described in detail below in the `patterns`_
subsection.

The ``match`` and ``case`` keywords are proposed to be soft keywords,
so that they are recognized as keywords at the beginning of a match
statement or case block respectively, but are allowed to be used in
other places as variable or argument names.

The proposed indentation structure is as following::

    match some_expression:
        case pattern_1:
            ...
        case pattern_2:
            ...

Here, `some_expression` represents the value that is being matched against,
which will be referred to hereafter as the *subject* of the match.


Match semantics
---------------

The proposed large scale semantics for choosing the match is to choose the first
matching pattern and execute the corresponding suite. The remaining patterns
are not tried. If there are no matching patterns, the statement 'falls
through', and execution continues at the following statement.

Essentially this is equivalent to a chain of ``if ... elif ... else``
statements. Note that unlike for the previously proposed ``switch`` statement,
the pre-computed dispatch dictionary semantics does not apply here.

There is no ``default`` or ``else`` case - instead the special wildcard
``_`` can be used (see the section on `capture_pattern`_) as a final
'catch-all' pattern.

Name bindings made during a successful pattern match outlive the executed suite
and can be used after the match statement. This follows the logic of other
Python statements that can bind names, such as ``for`` loop and ``with``
statement. For example::

  match shape:
      case Point(x, y):
          ...
      case Rectangle(x, y, _, _):
          ...
  print(x, y)  # This works

During failed pattern matches, some sub-patterns may succeed. For example,
while matching the value ``[0, 1, 2]`` with the pattern ``(0, x, 1)``, the
sub-pattern `x` may succeed if the list elements are matched from left to right.
The implementation may choose to either make persistent bindings for those
partial matches or not. User code including a `match` statement should not rely
on the bindings being made for a failed match, but also shouldn't assume that
variables are unchanged by a failed match. This part of the behavior is
left intentionally unspecified so different implementations can add
optimizations, and to prevent introducing semantic restrictions that could
limit the extensibility of this feature.

Note that some pattern types below define more specific rules about when
the binding is made.

.. _patterns:

Allowed patterns
----------------

We introduce the proposed syntax gradually. Here we start from the main
building blocks. The following patterns are supported:


.. _literal_pattern:

Literal Patterns
~~~~~~~~~~~~~~~~

Simplified syntax::

    literal_pattern:
        | number
        | string
        | 'None'
        | 'True'
        | 'False'


A literal pattern consists of a simple literal like a string, a number,
a Boolean literal (``True`` or ``False``), or ``None``::

  match number:
      case 0:
          print("Nothing")
      case 1:
          print("Just one")
      case 2:
          print("A couple")
      case -1:
          print("One less than nothing")
      case 1-1j:
          print("Good luck with that...")

Literal pattern uses equality with literal on the right hand side, so that
in the above example ``number == 0`` and then possibly ``number == 1``, etc
will be evaluated. Note that although technically negative numbers
are represented using unary minus, they are considered
literals for the purpose of pattern matching. Unary plus is not allowed.
Binary plus and minus are allowed only to join a real number and an imaginary
number to form a complex number, such as ``1+1j``.

Note that because equality (``__eq__``) is used, and the equivalency
between Booleans and the integers ``0`` and ``1``, there is no
practical difference between the following two::

      case True:
          ...

      case 1:
          ...

Triple-quoted strings are supported.  Raw strings and byte strings
are supported. F-strings are not allowed (since in general they are not
really literals).


.. _capture_pattern:

Capture Patterns
~~~~~~~~~~~~~~~~

Simplified syntax::

    capture_pattern: NAME

A capture pattern serves as an assignment target for the matched expression::

  match greeting:
      case "":
          print("Hello!")
      case name:
          print(f"Hi {name}!")

Only a single name is allowed (a dotted name is a constant value pattern).
A capture pattern always succeeds. A capture pattern appearing in a scope makes
the name local to that scope. For example, using ``name`` after the above
snippet may raise ``UnboundLocalError`` rather than ``NameError``, if
the ``""`` case clause was taken::

  match greeting:
      case "":
          print("Hello!")
      case name:
          print(f"Hi {name}!")
  if name == "Santa":      # <-- might raise UnboundLocalError
      ...                  # but works fine if greeting was not empty

While matching against each case clause, a name may be bound at most
once, having two capture patterns with coinciding names is an error::

  match data:
      case [x, x]:  # Error!
          ...

Note: one can still match on a collection with equal items using `guards`_.
Also, ``[x, y] | Point(x, y)`` is a legal pattern because the two
alternatives are never matched at the same time.

The single underscore (``_``) is not considered a ``NAME`` and treated specially
as a `wildcard pattern`_.

Reminder: ``None``, ``False`` and ``True`` are keywords denoting
literals, not names.

.. _wildcard_pattern:

Wildcard Pattern
~~~~~~~~~~~~~~~~

Simplified syntax::

    wildcard_pattern: "_"

The single underscore (``_``) name is a special kind of pattern that always
matches but *never* binds::

  match data:
      case [_, _]:
          print("Some pair")
          print(_)  # Error!

Given that no binding is made, it can be used as many times as desired, unlike
capture patterns.

.. _constant_value_pattern:

Constant Value Patterns
~~~~~~~~~~~~~~~~~~~~~~~

Simplified syntax::

    constant_pattern: NAME ('.' NAME)+

This is used to match against constants and enum values.
Every dotted name in a pattern is looked up using normal Python name
resolution rules, and the value is used for comparison by equality with
the match subject (same as for literals)::

  from enum import Enum

  class Sides(str, Enum):
      SPAM = "Spam"
      EGGS = "eggs"
      ...

  match entree[-1]:
      case Sides.SPAM:  # Compares entree[-1] == Sides.SPAM.
          response = "Have you got anything without Spam?"
      case side:  # Assigns side = entree[-1].
          response = f"Well, could I have their Spam instead of the {side} then?"

Note that there is no way to use unqualified names as constant value
patterns (they always denote variables to be captured).  See
`rejected ideas`_ for other syntactic alternatives that were
considered for constant value patterns.


.. _sequence_pattern:

Sequence Patterns
~~~~~~~~~~~~~~~~~

Simplified syntax::

  sequence_pattern:
      | '[' [values_pattern] ']'
      | '(' [value_pattern ',' [values pattern]] ')'
  values_pattern: ','.value_pattern+ ','?
  value_pattern: '*' capture_pattern | pattern

A sequence pattern follows the same semantics as unpacking assignment.
Like unpacking assignment, both tuple-like and list-like syntax can be
used, with identical semantics.  Each element can be an arbitrary
pattern; there may also be at most one ``*name`` pattern to catch all
remaining items::

  match collection:
      case 1, [x, *others]:
          print("Got 1 and a nested sequence")
      case (1, x):
          print(f"Got 1 and {x}")

To match a sequence pattern the subject must be an instance of
``collections.abc.Sequence``, and it cannot be any kind of string
(``str``, ``bytes``, ``bytearray``). It cannot be an iterator. For matching
on a specific collection class, see class pattern below.

The ``_`` wildcard can be starred to match sequences of varying lengths. For
example:

* ``[*_]`` matches a sequence of any length.
* ``(_, _, *_)``, matches any sequence of length two or more.
* ``["a", *_, "z"]`` matches any sequence of length two or more that starts with
  ``"a"`` and ends with ``"z"``.


.. _mapping_pattern:

Mapping Patterns
~~~~~~~~~~~~~~~~

Simplified syntax::

  mapping_pattern: '{' [items_pattern] '}'
  items_pattern: ','.key_value_pattern+ ','?
  key_value_pattern:
      | (literal_pattern | constant_pattern) ':' or_pattern
      | '**' capture_pattern


Mapping pattern is a generalization of iterable unpacking to mappings.
Its syntax is similar to dictionary display but each key and value are
patterns ``"{" (pattern ":" pattern)+ "}"``. A ``**name`` pattern is also
allowed, to extract the remaining items.  Only literal and constant value
patterns are allowed in key positions::

  import constants

  match config:
      case {"route": route}:
          process_route(route)
      case {constants.DEFAULT_PORT: sub_config, **rest}:
          process_config(sub_config, rest)

The subject must be an instance of ``collections.abc.Mapping``.
Extra keys in the subject are ignored even if ``**rest`` is not present.
This is different from sequence pattern, where extra items will cause a
match to fail. But mappings are actually different from sequences: they
have natural structural sub-typing behavior, i.e., passing a dictionary
with extra keys somewhere will likely just work.

For this reason, ``**_`` is invalid in mapping patterns; it would always be a
no-op that could be removed without consequence.

Matched key-value pairs must already be present in the mapping, and not created
on-the-fly by ``__missing__`` or ``__getitem__``.  For example,
``collections.defaultdict`` instances will only match patterns with keys that
were already present when the ``match`` block was entered.


.. _class_pattern:

Class Patterns
~~~~~~~~~~~~~~

Simplified syntax::

  class_pattern:
      | name_or_attr '(' ')'
      | name_or_attr '(' ','.pattern+ ','? ')'
      | name_or_attr '(' ','.keyword_pattern+ ','? ')'
      | name_or_attr '(' ','.pattern+ ',' ','.keyword_pattern+ ','? ')'
  keyword_pattern: NAME '=' or_pattern


A class pattern provides support for destructuring arbitrary objects.
There are two possible ways of matching on object attributes: by position
like ``Point(1, 2)``, and by name like ``Point(x=1, y=2)``. These
two can be combined, but positional match cannot follow a match by name.
Each item in a class pattern can be an arbitrary pattern. A simple
example::

  match shape:
      case Point(x, y):
          ...
      case Rectangle(x0, y0, x1, y1, painted=True):
          ...

Whether a match succeeds or not is determined by the equivalent of an
``isinstance`` call.  If the subject (``shape``, in the example) is not
an instance of the named class (``Point`` or ``Rectangle``), the match
fails.  Otherwise, it continues (see details in the `runtime`_
section).

The named class must inherit from ``type``.  It may be a single name
or a dotted name (e.g. ``some_mod.SomeClass`` or ``mod.pkg.Class``).
The leading name must not be ``_``, so e.g. ``_(...)`` and
``_.C(...)`` are invalid. Use ``object(foo=_)`` to check whether the
matched object has an attribute ``foo``.

By default, sub-patterns may only be matched by keyword for
user-defined classes. In order to support positional sub-patterns, a
custom ``__match_args__`` attribute is required.
The runtime allows matching against
arbitrarily nested patterns by chaining all of the instance checks and
attribute lookups appropriately.


Combining multiple patterns (OR patterns)
-----------------------------------------

Multiple alternative patterns can be combined into one using ``|``. This means
the whole pattern matches if at least one alternative matches.
Alternatives are tried from left to right and have short-circuit property,
subsequent patterns are not tried if one matched. Examples::

  match something:
      case 0 | 1 | 2:
          print("Small number")
      case [] | [_]:
          print("A short sequence")
      case str() | bytes():
          print("Something string-like")
      case _:
          print("Something else")

The alternatives may bind variables, as long as each alternative binds
the same set of variables (excluding ``_``).  For example::

  match something:
      case 1 | x:  # Error!
          ...
      case x | 1:  # Error!
          ...
      case one := [1] | two := [2]:  # Error!
          ...
      case Foo(arg=x) | Bar(arg=x):  # Valid, both arms bind 'x'
          ...
      case [x] | x:  # Valid, both arms bind 'x'
          ...


.. _guards:

Guards
------

Each *top-level* pattern can be followed by a **guard** of the form
``if expression``. A case clause succeeds if the pattern matches and the guard
evaluates to a true value. For example::

  match input:
      case [x, y] if x > MAX_INT and y > MAX_INT:
          print("Got a pair of large numbers")
      case x if x > MAX_INT:
          print("Got a large number")
      case [x, y] if x == y:
          print("Got equal items")
      case _:
          print("Not an outstanding input")

If evaluating a guard raises an exception, it is propagated onwards rather
than fail the case clause. Names that appear in a pattern are bound before the
guard succeeds. So this will work::

  values = [0]

  match values:
      case [x] if x:
          ...  # This is not executed
      case _:
          ...
  print(x)  # This will print "0"

Note that guards are not allowed for nested patterns, so that ``[x if x > 0]``
is a ``SyntaxError`` and ``1 | 2 if 3 | 4`` will be parsed as
``(1 | 2) if (3 | 4)``.


Walrus patterns
---------------

It is often useful to match a sub-pattern *and* bind the corresponding
value to a name. For example, it can be useful to write more efficient
matches, or simply to avoid repetition. To simplify such cases, any pattern
(other than the walrus pattern itself) can be preceded by a name and
the walrus operator (``:=``). For example::

  match get_shape():
      case Line(start := Point(x, y), end) if start == end:
          print(f"Zero length line at {x}, {y}")

The name on the left of the walrus operator can be used in a guard, in
the match suite, or after the match statement.  However, the name will
*only* be bound if the sub-pattern succeeds.  Another example::

  match group_shapes():
      case [], [point := Point(x, y), *other]:
          print(f"Got {point} in the second group")
          process_coordinates(x, y)
          ...

Technically, most such examples can be rewritten using guards and/or nested
match statements, but this will be less readable and/or will produce less
efficient code. Essentially, most of the arguments in PEP 572 apply here
equally.

The wildcard ``_`` is not a valid name here.


.. _runtime:

Runtime specification
=====================

The Match Protocol
------------------

The equivalent of an ``isinstance`` call is used to decide whether an
object matches a given class pattern and to extract the corresponding
attributes.  Classes requiring different matching semantics (such as
duck-typing) can do so by defining ``__instancecheck__`` (a
pre-existing metaclass hook) or by using ``typing.Protocol``.

The procedure is as following:

* The class object for ``Class`` in ``Class(<sub-patterns>)`` is
  looked up and ``isinstance(obj, Class)`` is called, where ``obj`` is
  the value being matched.  If false, the match fails.

* Otherwise, if any sub-patterns are given in the form of positional
  or keyword arguments, these are matched from left to right, as
  follows.  The match fails as soon as a sub-pattern fails; if all
  sub-patterns succeed, the overall class pattern match succeeds.

* If there are match-by-position items and the class has a
  ``__match_args__``, the item at position ``i``
  is matched against the value looked up by attribute
  ``__match_args__[i]``. For example, a pattern ``Point2D(5, 8)``,
  where ``Point2D.__match_args__ == ["x", "y"]``, is translated
  (approximately) into ``obj.x == 5 and obj.y == 8``.

* If there are more positional items than the length of
  ``__match_args__``, a ``TypeError`` is raised.

* If the ``__match_args__`` attribute is absent on the matched class,
  and one or more positional item appears in a match,
  ``TypeError`` is also raised. We don't fall back on
  using ``__slots__`` or ``__annotations__`` -- "In the face of ambiguity,
  refuse the temptation to guess."

* If there are any match-by-keyword items the keywords are looked up
  as attributes on the subject.  If the lookup succeeds the value is
  matched against the corresponding sub-pattern.  If the lookup fails,
  the match fails.

Such a protocol favors simplicity of implementation over flexibility and
performance. For other considered alternatives, see `extended matching`_.

For the most commonly-matched built-in types (``bool``,
``bytearray``, ``bytes``, ``dict``, ``float``,
``frozenset``, ``int``, ``list``, ``set``, ``str``, and ``tuple``), a
single positional sub-pattern is allowed to be passed to
the call. Rather than being matched against any particular attribute
on the subject, it is instead matched against the subject itself.  This
creates behavior that is useful and intuitive for these objects:

* ``bool(False)`` matches ``False`` (but not ``0``).
* ``tuple((0, 1, 2))`` matches ``(0, 1, 2)`` (but not ``[0, 1, 2]``).
* ``int(i)`` matches any ``int`` and binds it to the name ``i``.


Overlapping sub-patterns
------------------------

Certain classes of overlapping matches are detected at
runtime and will raise exceptions. In addition to basic checks
described in the previous subsection:

* The interpreter will check that two match items are not targeting the same
  attribute, for example ``Point2D(1, 2, y=3)`` is an error.

* It will also check that a mapping pattern does not attempt to match
  the same key more than once.


Special attribute ``__match_args__``
------------------------------------

The ``__match_args__`` attribute is always looked up on the type
object named in the pattern.  If present, it must be a list or tuple
of strings naming the allowed positional arguments.

In deciding what names should be available for matching, the
recommended practice is that class patterns should be the mirror of
construction; that is, the set of available names and their types
should resemble the arguments to ``__init__()``.

Only match-by-name will work by default, and classes should define
``__match_args__`` as a class attribute if they would like to support
match-by-position.  Additionally, dataclasses and named tuples will
support match-by-position out of the box. See below for more details.

Exceptions and side effects
---------------------------

While matching each case, the ``match`` statement may trigger execution of other
functions (for example ``__getitem__()``, ``__len__()`` or
a property). Almost every exception caused by those propagates outside of the
match statement normally. The only case where an exception is not propagated is
an ``AttributeError`` raised while trying to lookup an attribute while matching
attributes of a Class Pattern; that case results in just a matching failure,
and the rest of the statement proceeds normally.

The only side-effect carried on explicitly by the matching process is the binding of
names. However, the process relies on attribute access,
instance checks, ``len()``, equality and item access on the subject and some of
its components. It also evaluates constant value patterns and the left side of
class patterns. While none of those typically create any side-effects, some of
these objects could. This proposal intentionally leaves out any specification
of what methods are called or how many times. User code relying on that
behavior should be considered buggy.

The standard library
--------------------

To facilitate the use of pattern matching, several changes will be made to
the standard library:

* Namedtuples and dataclasses will have auto-generated ``__match_args__``.

* For dataclasses the order of attributes in the generated ``__match_args__``
  will be the same as the order of corresponding arguments in the generated
  ``__init__()`` method. This includes the situations where attributes are
  inherited from a superclass.

In addition, a systematic effort will be put into going through
existing standard library classes and adding ``__match_args__`` where
it looks beneficial.


.. _static checkers:

Static checkers specification
=============================

Exhaustiveness checks
---------------------

From a reliability perspective, experience shows that missing a case when
dealing with a set of possible data values leads to hard to debug issues,
thus forcing people to add safety asserts like this::

  def get_first(data: Union[int, list[int]]) -> int:
      if isinstance(data, list) and data:
          return data[0]
      elif isinstance(data, int):
          return data
      else:
          assert False, "should never get here"

PEP 484 specifies that static type checkers should support exhaustiveness in
conditional checks with respect to enum values. PEP 586 later generalized this
requirement to literal types.

This PEP further generalizes this requirement to
arbitrary patterns. A typical situation where this applies is matching an
expression with a union type::

  def classify(val: Union[int, Tuple[int, int], List[int]]) -> str:
      match val:
          case [x, y] if x > 0 and y > 0:
              return f"A pair of {x} and {y}"
          case [x, *other]:
              return f"A sequence starting with {x}"
          case int():
              return f"Some integer"
          # Type-checking error: some cases unhandled.

The exhaustiveness checks should also apply where both pattern matching
and enum values are combined::

  from enum import Enum
  from typing import Union

  class Level(Enum):
      BASIC = 1
      ADVANCED = 2
      PRO = 3

  class User:
      name: str
      level: Level

  class Admin:
      name: str

  account: Union[User, Admin]

  match account:
      case Admin(name=name) | User(name=name, level=Level.PRO):
          ...
      case User(level=Level.ADVANCED):
          ...
      # Type-checking error: basic user unhandled

Obviously, no ``Matchable`` protocol (in terms of PEP 544) is needed, since
every class is matchable and therefore is subject to the checks specified
above.


Sealed classes as algebraic data types
--------------------------------------

Quite often it is desirable to apply exhaustiveness to a set of classes without
defining ad-hoc union types, which is itself fragile if a class is missing in
the union definition. A design pattern where a group of record-like classes is
combined into a union is popular in other languages that support pattern
matching and is known under a name of algebraic data types [2]_.

We propose to add a special decorator class ``@sealed`` to the ``typing``
module [6]_, that will have no effect at runtime, but will indicate to static
type checkers that all subclasses (direct and indirect) of this class should
be defined in the same module as the base class.

The idea is that since all subclasses are known, the type checker can treat
the sealed base class as a union of all its subclasses. Together with
dataclasses this allows a clean and safe support of algebraic data types
in Python. Consider this example::

  from dataclasses import dataclass
  from typing import sealed

  @sealed
  class Node:
      ...

  class Expression(Node):
      ...

  class Statement(Node):
      ...

  @dataclass
  class Name(Expression):
      name: str

  @dataclass
  class Operation(Expression):
      left: Expression
      op: str
      right: Expression

  @dataclass
  class Assignment(Statement):
      target: str
      value: Expression

  @dataclass
  class Print(Statement):
      value: Expression

With such definition, a type checker can safely treat ``Node`` as
``Union[Name, Operation, Assignment, Print]``, and also safely treat e.g.
``Expression`` as ``Union[Name, Operation]``. So this will result in a type
checking error in the below snippet, because ``Name`` is not handled (and type
checker can give a useful error message)::

  def dump(node: Node) -> str:
      match node:
          case Assignment(target, value):
              return f"{target} = {dump(value)}"
          case Print(value):
              return f"print({dump(value)})"
          case Operation(left, op, right):
              return f"({dump(left)} {op} {dump(right)})"


Type erasure
------------

Class patterns are subject to runtime type erasure. Namely, although one
can define a type alias ``IntQueue = Queue[int]`` so that a pattern like
``IntQueue()`` is syntactically valid, type checkers should reject such a
match::

  queue: Union[Queue[int], Queue[str]]
  match queue:
      case IntQueue():  # Type-checking error here
          ...

Note that the above snippet actually fails at runtime with the current
implementation of generic classes in the ``typing`` module, as well as
with builtin generic classes in the recently accepted PEP 585, because
they prohibit ``isinstance`` checks.

To clarify, generic classes are not prohibited in general from participating
in pattern matching, just that their type parameters can't be explicitly
specified. It is still fine if sub-patterns or literals bind the type
variables. For example::

  from typing import Generic, TypeVar, Union

  T = TypeVar('T')

  class Result(Generic[T]):
      first: T
      other: list[T]

  result: Union[Result[int], Result[str]]

  match result:
      case Result(first=int()):
          ...  # Type of result is Result[int] here
      case Result(other=["foo", "bar", *rest]):
          ...  # Type of result is Result[str] here


Note about constants
--------------------

The fact that a capture pattern is always an assignment target may create unwanted
consequences when a user by mistake tries to "match" a value against
a constant instead of using the constant value pattern. As a result, at
runtime such match will always succeed and moreover override the value of
the constant. It is important therefore that static type checkers warn about
such situations. For example::

  from typing import Final

  MAX_INT: Final = 2 ** 64

  value = 0

  match value:
      case MAX_INT:  # Type-checking error here: cannot assign to final name
          print("Got big number")
      case _:
          print("Something else")


Precise type checking of star matches
-------------------------------------

Type checkers should perform precise type checking of star items in pattern
matching giving them either a heterogeneous ``list[T]`` type, or
a ``TypedDict`` type as specified by PEP 589. For example::

  stuff: Tuple[int, str, str, float]

  match stuff:
      case a, *b, 0.5:
          # Here a is int and b is list[str]
          ...


Performance Considerations
==========================

Ideally, a ``match`` statement should have good runtime performance compared
to an equivalent chain of if-statements. Although the history of programming
languages is rife with examples of new features which increased engineer
productivity at the expense of additional CPU cycles, it would be
unfortunate if the benefits of ``match`` were counter-balanced by a significant
overall decrease in runtime performance.

Although this PEP does not specify any particular implementation strategy,
a few words about the prototype implementation and how it attempts to
maximize performance are in order.

Basically, the prototype implementation transforms all of the ``match``
statement syntax into equivalent if/else blocks - or more accurately, into
Python byte codes that have the same effect. In other words, all of the
logic for testing instance types, sequence lengths, mapping keys and
so on are inlined in place of the ``match``.

This is not the only possible strategy, nor is it necessarily the best.
For example, the instance checks could be memoized, especially
if there are multiple instances of the same class type but with different
arguments in a single match statement. It is also theoretically
possible for a future implementation to process case clauses or sub-patterns in
parallel using a decision tree rather than testing them one by one.


Backwards Compatibility
=======================

This PEP is fully backwards compatible: the ``match`` and ``case``
keywords are proposed to be (and stay!) soft keywords, so their use as
variable, function, class, module or attribute names is not impeded at
all.

This is important because ``match`` is the name of a popular and
well-known function and method in the ``re`` module, which we have no
desire to break or deprecate.

The difference between hard and soft keywords is that hard keywords
are *always* reserved words, even in positions where they make no
sense (e.g. ``x = class + 1``), while soft keywords only get a special
meaning in context.  Since our parser backtracks, that means that on
different attempts to parse a code fragment it could interpret a soft
keyword differently.

For example, suppose the parser encounters the following input::

  match [x, y]:

The parser first attempts to parse this as an expression statement.
It interprets ``match`` as a NAME token, and then considers ``[x,
y]`` to be a double subscript.  It then encounters the colon and has
to backtrack, since an expression statement cannot be followed by a
colon.  The parser then backtracks to the start of the line and finds
that ``match`` is a soft keyword allowed in this position.  It then
considers ``[x, y]`` to be a list expression.  The colon then is just
what the parser expected, and the parse succeeds.


Impacts on third-party tools
============================

There are a lot of tools in the Python ecosystem that operate on Python
source code: linters, syntax highlighters, auto-formatters, and IDEs. These
will all need to be updated to include awareness of the ``match`` statement.

In general, these tools fall into one of two categories:

**Shallow** parsers don't try to understand the full syntax of Python, but
instead scan the source code for specific known patterns. IDEs, such as Visual
Studio Code, Emacs and TextMate, tend to fall in this category, since frequently
the source code is invalid while being edited, and a strict approach to parsing
would fail.

For these kinds of tools, adding knowledge of a new keyword is relatively
easy, just an addition to a table, or perhaps modification of a regular
expression.

**Deep** parsers understand the complete syntax of Python. An example of this
is the auto-formatter Black [9]_. A particular requirement with these kinds of
tools is that they not only need to understand the syntax of the current version
of Python, but older versions of Python as well.

The ``match`` statement uses a soft keyword, and it is one of the first major
Python features to take advantage of the capabilities of the new PEG parser. This
means that third-party parsers which are not 'PEG-compatible' will have a hard
time with the new syntax.

It has been noted that a number of these third-party tools leverage common parsing
libraries (Black for example uses a fork of the lib2to3 parser). It may be helpful
to identify widely-used parsing libraries (such as parso [10]_ and libCST [11]_)
and upgrade them to be PEG compatible.

However, since this work would need to be done not only for the match statement,
but for *any* new Python syntax that leverages the capabilities of the PEG parser,
it is considered out of scope for this PEP. (Although it is suggested that this
would make a fine Summer of Code project.)


Reference Implementation
========================

A `feature-complete CPython implementation
<https://github.com/brandtbucher/cpython/tree/patma>`_ is available on
GitHub.

An `interactive playground
<https://mybinder.org/v2/gh/gvanrossum/patma/master?urlpath=lab/tree/playground-622.ipynb>`_
based on the above implementation was created using Binder [12]_ and Jupyter [13]_.

Example Code
============

A small `collection of example code
<https://github.com/gvanrossum/patma/tree/master/examples>`_ is
available on GitHub.


.. _rejected ideas:

Rejected Ideas
==============

This general idea has been floating around for a pretty long time, and many
back and forth decisions were made. Here we summarize many alternative
paths that were taken but eventually abandoned.

Don't do this, pattern matching is hard to learn
------------------------------------------------

In our opinion, the proposed pattern matching is not more difficult than
adding ``isinstance()`` and ``getattr()`` to iterable unpacking. Also, we
believe the proposed syntax significantly improves readability for a wide
range of code patterns, by allowing to express *what* one wants to do, rather
than *how* to do it. We hope the few real code snippets we included in the PEP
above illustrate this comparison well enough. For more real code examples
and their translations see Ref. [7]_.


Don't do this, use existing method dispatching mechanisms
---------------------------------------------------------

We recognize that some of the use cases for the ``match`` statement overlap
with what can be done with traditional object-oriented programming (OOP) design
techniques using class inheritance. The ability to choose alternate
behaviors based on testing the runtime type of a match subject might
even seem heretical to strict OOP purists.

However, Python has always been a language that embraces a variety of
programming styles and paradigms. Classic Python design idioms such as
"duck"-typing go beyond the traditional OOP model.

We believe that there are important use cases where the use of ``match`` results
in a cleaner and more maintainable architecture. These use cases tend to
be characterized by a number of features:

* Algorithms which cut across traditional lines of data encapsulation. If an
  algorithm is processing heterogenous elements of different types (such as
  evaluating or transforming an abstract syntax tree, or doing algebraic
  manipulation of mathematical symbols), forcing the user to implement
  the algorithm as individual methods on each element type results in
  logic that is smeared across the entire codebase instead of being neatly
  localized in once place.
* Program architectures where the set of possible data types is relatively
  stable, but there is an ever-expanding set of operations to be performed
  on those data types. Doing this in a strict OOP fashion requires constantly
  adding new methods to both the base class and subclasses to support the new
  methods, "polluting" the base class with lots of very specialized method
  definitions, and causing widespread disruption and churn in the code. By
  contrast, in a ``match``-based dispatch, adding a new behavior merely
  involves writing a new ``match`` statement.
* OOP also does not handle dispatching based on the *shape* of an object, such
  as the length of a tuple, or the presence of an attribute -- instead any such
  dispatching decision must be encoded into the object's type. Shape-based
  dispatching is particularly interesting when it comes to handling "duck"-typed
  objects.

Where OOP is clearly superior is in the opposite case: where the set of possible
operations is relatively stable and well-defined, but there is an ever-growing
set of data types to operate on. A classic example of this is UI widget toolkits,
where there is a fixed set of interaction types (repaint, mouse click, keypress,
and so on), but the set of widget types is constantly expanding as developers
invent new and creative user interaction styles. Adding a new kind of widget
is a simple matter of writing a new subclass, whereas with a match-based approach
you end up having to add a new case clause to many widespread match statements.
We therefore don't recommend using ``match`` in such a situation.


Allow more flexible assignment targets instead
----------------------------------------------

There was an idea to instead just generalize the iterable unpacking to much
more general assignment targets, instead of adding a new kind of statement.
This concept is known in some other languages as "irrefutable matches". We
decided not to do this because inspection of real-life potential use cases
showed that in vast majority of cases destructuring is related to an ``if``
condition. Also many of those are grouped in a series of exclusive choices.


Make it an expression
---------------------

In most other languages pattern matching is represented by an expression, not
statement. But making it an expression would be inconsistent with other
syntactic choices in Python. All decision making logic is expressed almost
exclusively in statements, so we decided to not deviate from this.


Use a hard keyword
------------------

There were options to make ``match`` a hard keyword, or choose a different
keyword. Although using a hard keyword would simplify life for simple-minded
syntax highlighters, we decided not to use hard keyword for several reasons:

* Most importantly, the new parser doesn't require us to do this. Unlike with
  ``async`` that caused hardships with being a soft keyword for few releases,
  here we can make ``match`` a permanent soft keyword.

* ``match`` is so commonly used in existing code, that it would break almost
  every existing program and will put a burden to fix code on many people who
  may not even benefit from the new syntax.

* It is hard to find an alternative keyword that would not be commonly used
  in existing programs as an identifier, and would still clearly reflect the
  meaning of the statement.


Use ``as`` or ``|`` instead of ``case`` for case clauses
--------------------------------------------------------

The pattern matching proposed here is a combination of multi-branch control
flow (in line with ``switch`` in Algol-derived languages or ``cond`` in Lisp)
and object-deconstruction as found in functional languages.  While the proposed
keyword ``case`` highlights the multi-branch aspect, alternative keywords such
as ``as`` would equally be possible, highlighting the deconstruction aspect.
``as`` or ``with``, for instance, also have the advantage of already being
keywords in Python.  However, since ``case`` as a keyword can only occur as a
leading keyword inside  a ``match`` statement, it is easy for a parser to
distinguish between its use as a keyword or as a variable.

Other variants would use a symbol like ``|`` or ``=>``, or go entirely without
special marker.

Since Python is a statement-oriented language in the tradition of Algol, and as
each composite statement starts with an identifying keyword, ``case`` seemed to
be most in line with Python's style and traditions.


Use a flat indentation scheme
-----------------------------

There was an idea to use an alternative indentation scheme, for example where
every case clause would not be indented with respect to the initial ``match``
part::

  match expression:
  case pattern_1:
      ...
  case pattern_2:
      ...

The motivation is that although flat indentation saves some horizontal space,
it may look awkward to an eye of a Python programmer, because everywhere else
colon is followed by an indent. This will also complicate life for
simple-minded code editors. Finally, the horizontal space issue can be
alleviated by allowing "half-indent" (i.e. two spaces instead of four) for
match statements.

In sample programs using `match`, written as part of the development of this
PEP, a noticeable improvement in code brevity is observed, more than making up
for the additional indentation level.

Another proposal considered was to use flat indentation but put the
expression on the line after ``match:``, like this::

  match:
      expression
  case pattern_1:
      ...
  case pattern_2:
      ...

This was ultimately rejected because the first block would be a
novelty in Python's grammar: a block whose only content is a single
expression rather than a sequence of statements.


Alternatives for constant value pattern
---------------------------------------

This is probably the trickiest item. Matching against some pre-defined
constants is very common, but the dynamic nature of Python also makes it
ambiguous with capture patterns. Five other alternatives were considered:

* Use some implicit rules. For example if a name was defined in the global
  scope, then it refers to a constant, rather than representing a
  capture pattern::

    # Here, the name "spam" must be defined in the global scope (and
    # not shadowed locally). "side" must be local.

    match entree[-1]:
        case spam: ...  # Compares entree[-1] == spam.
        case side: ...  # Assigns side = entree[-1].

  This however can cause surprises and action at a distance if someone
  defines an unrelated coinciding name before the match statement.

* Use a rule based on the case of a name. In particular, if the name
  starts with a lowercase letter it would be a capture pattern, while if
  it starts with uppercase it would refer to a constant::

    match entree[-1]:
        case SPAM: ...  # Compares entree[-1] == SPAM.
        case side: ...  # Assigns side = entree[-1].

  This works well with the recommendations for naming constants from
  PEP 8. The main objection is that there's no other part of core
  Python where the case of a name is semantically significant.
  In addition, Python allows identifiers to use different scripts,
  many of which (e.g. CJK) don't have a case distinction.

* Use extra parentheses to indicate lookup semantics for a given name. For
  example::

    match entree[-1]:
        case (spam): ...  # Compares entree[-1] == spam.
        case side: ...    # Assigns side = entree[-1].

  This may be a viable option, but it can create some visual noise if used
  often. Also honestly it looks pretty unusual, especially in nested contexts.

  This also has the problem that we may want or need parentheses to
  disambiguate grouping in patterns, e.g. in ``Point(x, y=(y :=
  complex()))``.

* Introduce a special symbol, for example ``.``, ``$``, or ``^`` to
  indicate that a given name is a constant to be matched against, not
  to be assigned to.  An earlier version of this proposal used a
  leading-dot rule::

    match entree[-1]:
        case .spam: ...  # Compares entree[-1] == spam.
        case side: ...   # Assigns side = entree[-1].

  While potentially useful, it introduces strange-looking new syntax
  without making the pattern syntax any more expressive.  Indeed,
  named constants can be made to work with the existing rules by
  converting them to ``Enum`` types, or enclosing them in their own
  namespace (considered by the authors to be one honking great idea)::

    match entree[-1]:
        case Sides.SPAM: ...  # Compares entree[-1] == Sides.SPAM.
        case side: ...        # Assigns side = entree[-1].

  If needed, the leading-dot rule (or a similar variant) could be
  added back later with no backward-compatibility issues.

* There was also on idea to make lookup semantics the default, and require
  ``$`` or ``?`` to be used in capture patterns::

    match entree[-1]:
        case spam: ...   # Compares entree[-1] == spam.
        case side?: ...  # Assigns side = entree[-1].

  There are a few issues with this:

  * Capture patterns are more common in typical code, so it is
    undesirable to require special syntax for them.

  * The authors are not aware of any other language that adorns
    captures in this way.

  * None of the proposed syntaxes have any precedent in Python.

  * It would break the syntactic parallels of the current grammar::

      match coords:
          case ($x, $y):
              return Point(x, y)  # Why not "Point($x, $y)"?


In the end, these alternatives were rejected because of the mentioned drawbacks.


Disallow float literals in patterns
-----------------------------------

Because of the inexactness of floats, an early version of this proposal
did not allow floating-point constants to be used as match patterns. Part
of the justification for this prohibition is that Rust does this.

However, during implementation, it was discovered that distinguishing between
float values and other types required extra code in the VM that would slow
matches generally. Given that Python and Rust are very different languages
with different user bases and underlying philosophies, it was felt that
allowing float literals would not cause too much harm, and would be less
surprising to users.


Range matching patterns
-----------------------

This would allow patterns such as `1...6`. However, there are a host of
ambiguities:

* Is the range open, half-open, or closed? (I.e. is `6` included in the
  above example or not?)
* Does the range match a single number, or a range object?
* Range matching is often used for character ranges ('a'...'z') but that
  won't work in Python since there's no character data type, just strings.
* Range matching can be a significant performance optimization if you can
  pre-build a jump table, but that's not generally possible in Python due
  to the fact that names can be dynamically rebound.

Rather than creating a special-case syntax for ranges, it was decided
that allowing custom pattern objects (`InRange(0, 6)`) would be more flexible
and less ambiguous; however those ideas have been postponed for the time
being (See `deferred ideas`_).


Use dispatch dict semantics for matches
---------------------------------------

Implementations for classic ``switch`` statement sometimes use a pre-computed
hash table instead of a chained equality comparisons to gain some performance.
In the context of ``match`` statement this is technically also possible for
matches against literal patterns. However, having subtly different semantics
for different kinds of patterns would be too surprising for potentially
modest performance win.

We can still experiment with possible performance optimizations in this
direction if they will not cause semantic differences.


Use ``continue`` and ``break`` in case clauses.
-----------------------------------------------

Another rejected proposal was to define new meanings for ``continue``
and ``break`` inside of ``match``, which would have the following behavior:

* ``continue`` would exit the current case clause and continue matching
  at the next case clause.
* ``break`` would exit the match statement.

However, there is a serious drawback to this proposal: if the ``match`` statement
is nested inside of a loop, the meanings of ``continue`` and ``break`` are now
changed. This may cause unexpected behavior during refactorings; also, an
argument can be made that there are other means to get the same behavior (such
as using guard conditions), and that in practice it's likely that the existing
behavior of ``continue`` and ``break`` are far more useful.


AND (``&``) patterns
--------------------

This proposal defines an OR-pattern (``|``) to match one of several alternates;
why not also an AND-pattern (``&``)? Especially given that some other languages
(F# for example) support this.

However, it's not clear how useful this would be. The semantics for matching
dictionaries, objects and sequences already incorporates an implicit 'and': all
attributes and elements mentioned must be present for the match to succeed. Guard
conditions can also support many of the use cases that a hypothetical 'and'
operator would be used for.

In the end, it was decided that this would make the syntax more complex without
adding a significant benefit.


Negative match patterns
-----------------------

A negation of a match pattern using the operator ``!`` as a prefix would match
exactly if the pattern itself does not match.  For instance, ``!(3 | 4)``
would match anything except ``3`` or ``4``.

This was rejected because there is documented evidence [8]_ that this feature
is rarely useful (in languages which support it) or used as double negation
``!!`` to control variable scopes and prevent variable bindings (which does
not apply to Python). It can also be simulated using guard conditions.


Check exhaustiveness at runtime
-------------------------------

The question is what to do if no case clause has a matching pattern, and
there is no default case. An earlier version of the proposal specified that
the behavior in this case would be to throw an exception rather than
silently falling through.

The arguments back and forth were many, but in the end the EIBTI (Explicit
Is Better Than Implicit) argument won out: it's better to have the programmer
explicitly throw an exception if that is the behavior they want.

For cases such as sealed classes and enums, where the patterns are all known
to be members of a discrete set, `static checkers`_ can warn about missing
patterns.


Type annotations for pattern variables
--------------------------------------

The proposal was to combine patterns with type annotations::

  match x:
      case [a: int, b: str]: print(f"An int {a} and a string {b}:)
      case [a: int, b: int, c: int]: print(f"Three ints", a, b, c)
      ...

This idea has a lot of problems. For one, the colon can only
be used inside of brackets or parens, otherwise the syntax becomes
ambiguous. And because Python disallows ``isinstance()`` checks
on generic types, type annotations containing generics will not
work as expected.


Allow ``*rest`` in class patterns
---------------------------------

It was proposed to allow ``*rest`` in a class pattern, giving a
variable to be bound to all positional arguments at once (similar to
its use in unpacking assignments).  It would provide some symmetry
with sequence patterns.  But it might be confused with a feature to
provide the *values* for all positional arguments at once.  And there
seems to be no practical need for it, so it was scrapped.  (It could
easily be added at a later stage if a need arises.)

Disallow ``_.a`` in constant value patterns
------------------------------------------------------

The first public draft said that the initial name in a constant value
pattern must not be ``_`` because ``_`` has a special meaning in
pattern matching, so this would be invalid::

    case _.a: ...

(However, ``a._`` would be legal and load the attribute with name
``_`` of the object ``a`` as usual.)

There was some pushback against this on python-dev (some people have a
legitimate use for ``_`` as an important global variable, esp. in
i18n) and the only reason for this prohibition was to prevent some
user confusion.  But it's not the hill to die on.

Use some other token as wildcard
--------------------------------

It has been proposed to use ``...`` (i.e., the ellipsis token) or
``*`` (star) as a wildcard.  However, both these look as if an
arbitrary number of items is omitted::

    case [a, ..., z]: ...
    case [a, *, z]: ...

Both look like the would match a sequence of at two or more items,
capturing the first and last values.

In addition, if ``*`` were to be used as the wildcard character, we
would have to come up with some other way to capture the rest of a
sequence, currently spelled like this::

    case [first, second, *rest]: ...

Using an ellipsis would also be more confusing in documentation and
examples, where ``...`` is routinely used to indicate something
obvious or irrelevant.  (Yes, this would also be an argument against
the other uses of ``...`` in Python, but that water is already under
the bridge.)

Another proposal was to use ``?``.  This could be acceptable, although
it would require modifying the tokenizer.

Also, ``_`` is already used
as a throwaway target in other contexts, and this use is pretty
similar.  This example is from ``difflib.py`` in the stdlib::

  for tag, _, _, j1, j2 in group: ...

Perhaps the most convincing argument is that ``_`` is used as the
wildcard in every other language we've looked at supporting pattern
matching: C#, Elixir, Erlang, F#, Haskell, Mathematica, OCaml, Ruby,
Rust, Scala, and Swift.  Now, in general, we should not be concerned
too much with what another language does, since Python is clearly
different from all these languages.  However, if there is such an
overwhelming and strong consensus, Python should not go out of its way
to do something completely different -- particularly given that ``_``
works well in Python and is already in use as a throwaway target.

Note that ``_`` is not assigned to by patterns -- this avoids
conflicts with the use of ``_`` as a marker for translatable strings
and an alias for ``gettext.gettext``, as recommended by the
``gettext`` module documentation.

Use some other syntax instead of ``|`` for OR patterns
------------------------------------------------------

A few alternatives to using ``|`` to separate the alternatives in OR
patterns have been proposed.  Instead of::

  case 401|403|404:
      print("Some HTTP error")

the following proposals have been fielded:

- Use a comma::

    case 401, 403, 404:
      print("Some HTTP error")

  This looks too much like a tuple -- we would have to find a
  different way to spell tuples, and the construct would have to be
  parenthesized inside the argument list of a class pattern.  In
  general, commas already have many different meanings in Python, we
  shouldn't add more.

- Allow stacked cases::

    case 401:
    case 403:
    case 404:
      print("Some HTTP error")

  This is how this would be done in C, using its fall-through
  semantics for cases.  However, we don't want to mislead people into
  thinking that ``match``/``case`` uses fall-through semantics (which
  are a common source of bugs in C).  Also, this would be a novel
  indentation pattern, which might make it harder to support in IDEs
  and such (it would break the simple rule "add an indentation level
  after a line ending in a colon").  Finally, this wouldn't support
  OR patterns nested inside other patterns.

- Use ``case in`` followed by a comma-separated list::

    case in 401, 403, 404:
      print("Some HTTP error")

  This wouldn't work for OR patterns nested inside other patterns,
  like::

    case Point(0|1, 0|1):
        print("A corner of the unit square")

- Use the ``or`` keyword::

    case 401 or 403 or 404:
        print("Some HTTP error")

  This could work, and the readability is not too different from using
  ``|``.  Some users expressed a preference for ``or`` because they
  associate ``|`` with bitwise OR.  However:

  1. Many other languages that have pattern matching use ``|`` (the
     list includes Elixir, Erlang, F#, Mathematica, OCaml, Ruby, Rust,
     and Scala).
  2. ``|`` is shorter, which may contribute to the readability of
     nested patterns like ``Point(0|1, 0|1)``.
  3. Some people mistakenly believe that ``|`` has the wrong priority;
     but since patterns don't support other operators it has the same
     priority as in expressions.
  4. Python users use ``or`` very frequently, and may build an
     impression that it is strongly associated with Boolean
     short-circuiting.
  5. ``|`` is used between alternatives in regular expressions
     and in EBNF grammars (like Python's own).
  6. ``|`` not just used for bitwise OR -- it's used for set unions,
     dict merging (:pep:`584`) and is being considered as an
     alternative to ``typing.Union`` (:pep:`604`).
  7. ``|`` works better as a visual separator, especially between
     strings.  Compare::

       case "spam" or "eggs" or "cheese":

     to::

       case "spam" | "eggs" | "cheese":

Add an ``else`` clause
----------------------

We decided not to add an ``else`` clause for several reasons.

- It is redundant, since we already have ``case _:``

- There will forever be confusion about the indentation level of the
  ``else:`` -- should it align with the list of cases or with the
  ``match`` keyword?

- Completionist arguments like "every other statement has one" are
  false -- only those statements have an ``else`` clause where it adds
  new functionality.


.. _deferred ideas:

Deferred Ideas
==============

There were a number of proposals to extend the matching syntax that we
decided to postpone for possible future PEP. These fall into the realm of
"cool idea but not essential", and it was felt that it might be better to
acquire some real-world data on how the match statement will be used in
practice before moving forward with some of these proposals.

Note that in each case, the idea was judged to be a "two-way door",
meaning that there should be no backwards-compatibility issues with adding
these features later.

One-off syntax variant
----------------------

While inspecting some code-bases that may benefit the most from the proposed
syntax, it was found that single clause matches would be used relatively often,
mostly for various special-casing. In other languages this is supported in
the form of one-off matches. We proposed to support such one-off matches too::

  if match value as pattern [and guard]:
      ...

or, alternatively, without the ``if``::

  match value as pattern [if guard]:
      ...

as equivalent to the following expansion::

  match value:
      case pattern [if guard]:
          ...

To illustrate how this will benefit readability, consider this (slightly
simplified) snippet from real code::

  if isinstance(node, CallExpr):
      if (isinstance(node.callee, NameExpr) and len(node.args) == 1 and
              isinstance(node.args[0], NameExpr)):
          call = node.callee.name
          arg = node.args[0].name
          ...  # Continue special-casing 'call' and 'arg'
  ...  # Follow with common code

This can be rewritten in a more straightforward way as::

  if match node as CallExpr(callee=NameExpr(name=call), args=[NameExpr(name=arg)]):
      ...  # Continue special-casing 'call' and 'arg'
  ...  # Follow with common code

This one-off form would not allow ``elif match`` statements, as it was only
meant to handle a single pattern case. It was intended to be special case
of a ``match`` statement, not a special case of an ``if`` statement::

  if match value_1 as patter_1 [and guard_1]:
      ...
  elif match value_2 as pattern_2 [and guard_2]:  # Not allowed
      ...
  elif match value_3 as pattern_3 [and guard_3]:  # Not allowed
      ...
  else:  # Also not allowed
      ...

This would defeat the purpose of one-off matches as a complement to exhaustive
full matches - it's better and clearer to use a full match in this case.

Similarly, ``if not match`` would not be allowed, since ``match ... as ...`` is not
an expression. Nor do we propose a ``while match`` construct present in some languages
with pattern matching, since although it may be handy, it will likely be used
rarely.

Other pattern-based constructions
---------------------------------

Many other languages supporting pattern-matching use it as a basis for multiple
language constructs, including a matching operator, a generalized form
of assignment, a filter for loops, a method for synchronizing communication,
or specialized if statements. Some of these were mentioned in the discussion
of the first draft. Another question asked was why this particular form (joining
binding and conditional selection) was chosen while other forms were not.

Introducing more uses of patterns would be too bold and premature given the
experience we have using patterns, and would make this proposal too
complicated. The statement as presented provides a form of the feature that
is sufficiently general to be useful while being self-contained, and without
having a massive impact on the syntax and semantics of the language as a whole.

After some experience with this feature, the community may have a better
feeling for what other uses of pattern matching could be valuable in Python.

Algebraic matching of repeated names
------------------------------------

A technique occasionally seen in functional languages like Erlang and Elixir is
to use a match variable multiple times in the same pattern::

  match value:
      case Point(x, x):
          print("Point is on a diagonal!")

The idea here is that the first appearance of ``x`` would bind the value
to the name, and subsequent occurrences would verify that the incoming
value was equal to the value previously bound. If the value was not equal,
the match would fail.

However, there are a number of subtleties involved with mixing load-store
semantics for capture patterns. For the moment, we decided to make repeated
use of names within the same pattern an error; we can always relax this
restriction later without affecting backwards compatibility.

Note that you **can** use the same name more than once in alternate choices::

  match value:
      case x | [x]:
          # etc.


.. _extended matching:

Custom matching protocol
------------------------

During the initial design discussions for this PEP, there were a lot of ideas
thrown around about custom matchers. There were a couple of motivations for
this:

* Some classes might want to expose a different set of "matchable" names
  than the actual class properties.
* Some classes might have properties that are expensive to calculate, and
  therefore shouldn't be evaluated unless the match pattern actually needed
  access to them.
* There were ideas for exotic matchers such as ``IsInstance()``,
  ``InRange()``, ``RegexMatchingGroup()`` and so on.
* In order for built-in types and standard library classes to be able
  to support matching in a reasonable and intuitive way, it was believed
  that these types would need to implement special matching logic.

These customized match behaviors would be controlled by a special
``__match__`` method on the class name. There were two competing variants:

* A 'full-featured' match protocol which would pass in not only
  the subject to be matched, but detailed information about
  which attributes the specified pattern was interested in.
* A simplified match protocol, which only passed in the subject value,
  and which returned a "proxy object" (which in most cases could be
  just the subject) containing the matchable attributes.

Here's an example of one version of the more complex protocol proposed::

  match expr:
      case BinaryOp(left=Number(value=x), op=op, right=Number(value=y)):
          ...

  from types import PatternObject
  BinaryOp.__match__(
      (),
      {
          "left": PatternObject(Number, (), {"value": ...}, -1, False),
          "op": ...,
          "right": PatternObject(Number, (), {"value": ...}, -1, False),
      },
      -1,
      False,
  )

One drawback of this protocol is that the arguments to ``__match__``
would be expensive to construct, and could not be pre-computed due to
the fact that, because of the way names are bound, there are no real
constants in Python. It also meant that the ``__match__`` method would
have to re-implement much of the logic of matching which would otherwise
be implemented in C code in the Python VM. As a result, this option would
perform poorly compared to an equilvalent ``if``-statement.

The simpler protocol suffered from the fact that although it was more
performant, it was much less flexible, and did not allow for many of
the creative custom matchers that people were dreaming up.

Late in the design process, however, it was realized that the need for
a custom matching protocol was much less than anticipated. Virtually
all the realistic (as opposed to fanciful) uses cases brought up could
be handled by the built-in matching behavior, although in a few cases
an extra guard condition was required to get the desired effect.

Moreover, it turned out that none of the standard library classes really
needed any special matching support other than an appropriate
``__match_args__`` property.

The decision to postpone this feature came with a realization that this is
not a one-way door; that a more flexible and customizable matching protocol
can be added later, especially as we gain more experience with real-world
use cases and actual user needs.

The authors of this PEP expect that the ``match`` statement will evolve
over time as usage patterns and idioms evolve, in a way similar to what
other "multi-stage" PEPs have done in the past. When this happens, the
extended matching issue can be revisited.


Parameterized Matching Syntax
-----------------------------

(Also known as "Class Instance Matchers".)

This is another variant of the "custom match classes" idea that would allow
diverse kinds of custom matchers mentioned in the previous section -- however,
instead of using an extended matching protocol, it would be achieved by
introducing an additional pattern type with its own syntax. This pattern type
would accept two distinct sets of parameters: one set which consists of the
actual parameters passed into the pattern object's constructor, and another
set representing the binding variables for the pattern.

The ``__match__`` method of these objects could use the constructor parameter
values in deciding what was a valid match.

This would allow patterns such as ``InRange<0, 6>(value)``, which would match
a number in the range 0..6 and assign the matched value to 'value'. Similarly,
one could have a pattern which tests for the existence of a named group in
a regular expression match result (different meaning of the word 'match').

Although there is some support for this idea, there was a lot of bikeshedding
on the syntax (there are not a lot of attractive options available)
and no clear consensus was reached, so it was decided that for now, this
feature is not essential to the PEP.


Pattern Utility Library
-----------------------

Both of the previous ideas would be accompanied by a new Python standard
library module which would contain a rich set of exotic and useful matchers.
However, it it not really possible to implement such a library without
adopting one of the extended pattern proposals given in the previous sections,
so this idea is also deferred.


Acknowledgments
===============

We are grateful for the help of the following individuals (among many
others) for helping out during various phases of the writing of this
PEP:

- Gregory P. Smith
- Jim Jewett
- Mark Shannon
- Nate Lust
- Taine Zhao


Version History
===============

1. Initial version

2. Substantial rewrite, including:

   - Minor clarifications, grammar and typo corrections
   - Rename various concepts
   - Additional discussion of rejected ideas, including:

     - Why we choose ``_`` for wildcard patterns
     - Why we choose ``|`` for OR patterns
     - Why we choose not to use special syntax for capture variables
     - Why this pattern matching operation and not others

   - Clarify exception and side effect semantics
   - Clarify partial binding semantics
   - Drop restriction on use of ``_`` in load contexts
   - Drop the default single positional argument being the whole
     subject except for a handful of built-in types
   - Simplify behavior of ``__match_args__``
   - Drop the ``__match__`` protocol (moved to `deferred ideas`_)
   - Drop ``ImpossibleMatchError`` exception
   - Drop leading dot for loads (moved to `deferred ideas`_)
   - Reworked the initial sections (everything before `syntax`_)
   - Added an overview of all the types of patterns before the
     detailed description
   - Added simplified syntax next to the description of each pattern
   - Separate description of the wildcard from capture patterns
   - Added Daniel F Moisset as sixth co-author

References
==========

.. [1]
   https://en.wikipedia.org/wiki/Pattern_matching

.. [2]
   https://en.wikipedia.org/wiki/Algebraic_data_type

.. [3]
   https://doc.rust-lang.org/reference/patterns.html

.. [4]
   https://docs.scala-lang.org/tour/pattern-matching.html

.. [5]
   https://docs.python.org/3/library/dataclasses.html

.. [6]
   https://docs.python.org/3/library/typing.html

.. [7]
   https://github.com/gvanrossum/patma/blob/master/EXAMPLES.md

.. [8]
   https://dl.acm.org/doi/abs/10.1145/2480360.2384582

.. [9]
   https://black.readthedocs.io/en/stable/

.. [10]
   https://github.com/davidhalter/parso

.. [11]
   https://github.com/Instagram/LibCST

.. [12]
   https://mybinder.org

.. [13]
   https://jupyter.org


.. _Appendix A:

Appendix A -- Full Grammar
==========================

Here is the full grammar for ``match_stmt``.  This is an additional
alternative for ``compound_stmt``.  It should be understood that
``match`` and ``case`` are soft keywords, i.e. they are not reserved
words in other grammatical contexts (including at the start of a line
if there is no colon where expected).  By convention, hard keywords
use single quotes while soft keywords use double quotes.

Other notation used beyond standard EBNF:

- ``SEP.RULE+`` is shorthand for ``RULE (SEP RULE)*``
- ``!RULE`` is a negative lookahead assertion

::

  match_expr:
      | star_named_expression ',' star_named_expressions?
      | named_expression
  match_stmt: "match" match_expr ':' NEWLINE INDENT case_block+ DEDENT
  case_block: "case" patterns [guard] ':' block
  guard: 'if' named_expression
  patterns: value_pattern ',' [values_pattern] | pattern
  pattern: walrus_pattern | or_pattern
  walrus_pattern: NAME ':=' or_pattern
  or_pattern: '|'.closed_pattern+
  closed_pattern:
      | capture_pattern
      | literal_pattern
      | constant_pattern
      | group_pattern
      | sequence_pattern
      | mapping_pattern
      | class_pattern
  capture_pattern: NAME !('.' | '(' | '=')
  literal_pattern:
      | signed_number !('+' | '-')
      | signed_number '+' NUMBER
      | signed_number '-' NUMBER
      | strings
      | 'None'
      | 'True'
      | 'False'
  constant_pattern: attr !('.' | '(' | '=')
  group_pattern: '(' patterns ')'
  sequence_pattern: '[' [values_pattern] ']' | '(' ')'
  mapping_pattern: '{' items_pattern? '}'
  class_pattern:
      | name_or_attr '(' ')'
      | name_or_attr '(' ','.pattern+ ','? ')'
      | name_or_attr '(' ','.keyword_pattern+ ','? ')'
      | name_or_attr '(' ','.pattern+ ',' ','.keyword_pattern+ ','? ')'
  signed_number: NUMBER | '-' NUMBER
  attr: name_or_attr '.' NAME
  name_or_attr: attr | NAME
  values_pattern: ','.value_pattern+ ','?
  items_pattern: ','.key_value_pattern+ ','?
  keyword_pattern: NAME '=' or_pattern
  value_pattern: '*' capture_pattern | pattern
  key_value_pattern:
      | (literal_pattern | constant_pattern) ':' or_pattern
      | '**' capture_pattern


Copyright
=========

This document is placed in the public domain or under the
CC0-1.0-Universal license, whichever is more permissive.



..
   Local Variables:
   mode: indented-text
   indent-tabs-mode: nil
   sentence-end-double-space: t
   fill-column: 70
   coding: utf-8
   End: