How to check if a string is a valid python identifier? including keyword check?

Question

Does anyone know if there is any builtin python method that will check if something is a valid python variable name, INCLUDING a check against reserved keywords? (so, ie, something like 'in' or 'for' would fail...)

Failing that, does anyone know of where I can get a list of reserved keywords (ie, dyanamically, from within python, as opposed to copy-and-pasting something from the online docs)? Or, have another good way of writing your own check?

Surprisingly, testing by wrapping a setattr in try/except doesn't work, as something like this:

setattr(myObj, 'My Sweet Name!', 23)

...actually works! (...and can even be retrieved with getattr!)

Answer 1

Python 3

Python 3 now has 'foo'.isidentifier(), so that seems to be the best solution for recent Python versions (thanks fellow runciter@freenode for suggestion). However, somewhat counter-intuitively, it does not check against the list of keywords, so combination of both must be used:

import keyword

def isidentifier(ident: str) -> bool:
    """Determines if string is valid Python identifier."""

    if not isinstance(ident, str):
        raise TypeError("expected str, but got {!r}".format(type(ident)))

    if not ident.isidentifier():
        return False

    if keyword.iskeyword(ident):
        return False

    return True

Python 2

For Python 2, easiest possible way to check if given string is valid Python identifier is to let Python parse it itself.

There are two possible approaches. Fastest is to use ast, and check if AST of single expression is of desired shape:

import ast

def isidentifier(ident):
    """Determines, if string is valid Python identifier."""

    # Smoke test — if it's not string, then it's not identifier, but we don't
    # want to just silence exception. It's better to fail fast.
    if not isinstance(ident, str):
        raise TypeError("expected str, but got {!r}".format(type(ident)))

    # Resulting AST of simple identifier is <Module [<Expr <Name "foo">>]>
    try:
        root = ast.parse(ident)
    except SyntaxError:
        return False

    if not isinstance(root, ast.Module):
        return False

    if len(root.body) != 1:
        return False

    if not isinstance(root.body[0], ast.Expr):
        return False

    if not isinstance(root.body[0].value, ast.Name):
        return False

    if root.body[0].value.id != ident:
        return False

    return True

Another is to let tokenize module split the identifier into the stream of tokens, and check it only contains our name:

import keyword
import tokenize

def isidentifier(ident):
    """Determines if string is valid Python identifier."""

    # Smoke test - if it's not string, then it's not identifier, but we don't
    # want to just silence exception. It's better to fail fast.
    if not isinstance(ident, str):
        raise TypeError("expected str, but got {!r}".format(type(ident)))

    # Quick test - if string is in keyword list, it's definitely not an ident.
    if keyword.iskeyword(ident):
        return False

    readline = lambda g=(lambda: (yield ident))(): next(g)
    tokens = list(tokenize.generate_tokens(readline))

    # You should get exactly 2 tokens
    if len(tokens) != 2:
        return False

    # First is NAME, identifier.
    if tokens[0][0] != tokenize.NAME:
        return False

    # Name should span all the string, so there would be no whitespace.
    if ident != tokens[0][1]:
        return False

    # Second is ENDMARKER, ending stream
    if tokens[1][0] != tokenize.ENDMARKER:
        return False

    return True

The same function, but compatible with Python 3, looks like this:

import keyword
import tokenize

def isidentifier_py3(ident):
    """Determines if string is valid Python identifier."""

    # Smoke test — if it's not string, then it's not identifier, but we don't
    # want to just silence exception. It's better to fail fast.
    if not isinstance(ident, str):
        raise TypeError("expected str, but got {!r}".format(type(ident)))

    # Quick test — if string is in keyword list, it's definitely not an ident.
    if keyword.iskeyword(ident):
        return False

    readline = lambda g=(lambda: (yield ident.encode('utf-8-sig')))(): next(g)
    tokens = list(tokenize.tokenize(readline))

    # You should get exactly 3 tokens
    if len(tokens) != 3:
        return False

    # If using Python 3, first one is ENCODING, it's always utf-8 because 
    # we explicitly passed in UTF-8 BOM with ident.
    if tokens[0].type != tokenize.ENCODING:
        return False

    # Second is NAME, identifier.
    if tokens[1].type != tokenize.NAME:
        return False

    # Name should span all the string, so there would be no whitespace.
    if ident != tokens[1].string:
        return False

    # Third is ENDMARKER, ending stream
    if tokens[2].type != tokenize.ENDMARKER:
        return False

    return True

However, be aware of bugs in Python 3 tokenize implementation that reject some completely valid identifiers like ℘᧚, ﮯ and 贈ᩭ. ast works fine though. Generally, I'd advise against using tokenize-based implemetation for actual checks.

Also, some may consider heavy machinery like AST parser to be a tad overkill. This simple implementation is self-contained and guaranteed to work on any Python 2:

import keyword
import string

def isidentifier(ident):
    """Determines if string is valid Python identifier."""

    if not isinstance(ident, str):
        raise TypeError("expected str, but got {!r}".format(type(ident)))

    if not ident:
        return False

    if keyword.iskeyword(ident):
        return False

    first = '_' + string.lowercase + string.uppercase
    if ident[0] not in first:
        return False

    other = first + string.digits
    for ch in ident[1:]:
        if ch not in other:
            return False

    return True

Here are few tests to check these all work:

assert(isidentifier('foo'))
assert(isidentifier('foo1_23'))
assert(not isidentifier('pass'))    # syntactically correct keyword
assert(not isidentifier('foo '))    # trailing whitespace
assert(not isidentifier(' foo'))    # leading whitespace
assert(not isidentifier('1234'))    # number
assert(not isidentifier('1234abc')) # number and letters
assert(not isidentifier(''))      # Unicode not from allowed range
assert(not isidentifier(''))        # empty string
assert(not isidentifier('   '))     # whitespace only
assert(not isidentifier('foo bar')) # several tokens
assert(not isidentifier('no-dashed-names-for-you')) # no such thing in Python

# Unicode identifiers are only allowed in Python 3:
assert(isidentifier('℘᧚')) # Unicode $Other_ID_Start and $Other_ID_Continue

Performance

All measurements has been conducted on my machine (MBPr Mid 2014) on the same randomly generated test set of 1 500 000 elements, 1000 000 valid and 500 000 invalid. YMMV

== Python 3:
method | calls/sec | faster
---------------------------
token  |    48 286 |  1.00x
ast    |   175 530 |  3.64x
native | 1 924 680 | 39.86x

== Python 2:
method | calls/sec | faster
---------------------------
token  |    83 994 |  1.00x
ast    |   208 206 |  2.48x
simple | 1 066 461 | 12.70x

Answer 2

The keyword module contains the list of all reserved keywords:

>>> import keyword
>>> keyword.iskeyword("in")
True
>>> keyword.kwlist
['and', 'as', 'assert', 'break', 'class', 'continue', 'def', 'del', 'elif', 'else', 'except', 'exec', 'finally', 'for', 'from', 'global', 'if', 'import', 'in', 'is', 'lambda', 'not', 'or', 'pass', 'print', 'raise', 'return', 'try', 'while', 'with', 'yield']

Note that this list will be different depending on what major version of Python you are using, as the list of keywords changes (especially between Python 2 and Python 3).

If you also want all builtin names, use __builtins__

>>> dir(__builtins__)
['ArithmeticError', 'AssertionError', 'AttributeError', 'BaseException', 'BlockingIOError', 'BrokenPipeError', 'BufferError', 'BytesWarning', 'ChildProcessError', 'ConnectionAbortedError', 'ConnectionError', 'ConnectionRefusedError', 'ConnectionResetError', 'DeprecationWarning', 'EOFError', 'Ellipsis', 'EnvironmentError', 'Exception', 'False', 'FileExistsError', 'FileNotFoundError', 'FloatingPointError', 'FutureWarning', 'GeneratorExit', 'IOError', 'ImportError', 'ImportWarning', 'IndentationError', 'IndexError', 'InterruptedError', 'IsADirectoryError', 'KeyError', 'KeyboardInterrupt', 'LookupError', 'MemoryError', 'NameError', 'None', 'NotADirectoryError', 'NotImplemented', 'NotImplementedError', 'OSError', 'OverflowError', 'PendingDeprecationWarning', 'PermissionError', 'ProcessLookupError', 'ReferenceError', 'ResourceWarning', 'RuntimeError', 'RuntimeWarning', 'StopIteration', 'SyntaxError', 'SyntaxWarning', 'SystemError', 'SystemExit', 'TabError', 'TimeoutError', 'True', 'TypeError', 'UnboundLocalError', 'UnicodeDecodeError', 'UnicodeEncodeError', 'UnicodeError', 'UnicodeTranslateError', 'UnicodeWarning', 'UserWarning', 'ValueError', 'Warning', 'ZeroDivisionError', '_', '__build_class__', '__debug__', '__doc__', '__import__', '__name__', '__package__', 'abs', 'all', 'any', 'ascii', 'bin', 'bool', 'bytearray', 'bytes', 'callable', 'chr', 'classmethod', 'compile', 'complex', 'copyright', 'credits', 'delattr', 'dict', 'dir', 'divmod', 'enumerate', 'eval', 'exec', 'exit', 'filter', 'float', 'format', 'frozenset', 'getattr', 'globals', 'hasattr', 'hash', 'help', 'hex', 'id', 'input', 'int', 'isinstance', 'issubclass', 'iter', 'len', 'license', 'list', 'locals', 'map', 'max', 'memoryview', 'min', 'next', 'object', 'oct', 'open', 'ord', 'pow', 'print', 'property', 'quit', 'range', 'repr', 'reversed', 'round', 'set', 'setattr', 'slice', 'sorted', 'staticmethod', 'str', 'sum', 'super', 'tuple', 'type', 'vars', 'zip']

And note that some of these (like copyright) are not really that big of a deal to override.

One more caveat: note that in Python 2, True, False, and None are not considered keywords. However, assigning to None is a SyntaxError. Assigning to True or False is allowed, though not recommended (same with any other builtin). In Python 3, they are keywords, so this is not an issue.

Answer 3

John: as a slight improvement, I added a $ in the re, otherwise, the test does not detect spaces:

import keyword 
import re
my_var = "$testBadVar"
print re.match("[_A-Za-z][_a-zA-Z0-9]*$",my_var) and not keyword.iskeyword(my_var)

Answer 4

The list of python keywords is short so you can just check syntax with a simple regex and membership in a relatively small list of keywords

import keyword #thanks asmeurer
import re
my_var = "$testBadVar"
print re.match("[_A-Za-z][_a-zA-Z0-9]*",my_var) and not keyword.iskeyword(my_var)

a shorter but more dangerous alternative would be

my_bad_var="%#ASD"
try:exec("{0}=1".format(my_bad_var))
except SyntaxError: #this maynot be right error
   print "Invalid variable name!"

and lastly a slightly safer variant

my_bad_var="%#ASD"

try:
  cc = compile("{0}=1".format(my_bad_var),"asd","single")
  eval(cc)
  print "VALID"
 except SyntaxError: #maybe different error
  print "INVALID!"

Answer 5

I needed to check for Python 3 identifiers from Python 2 code. I used a regex based on the docs:

import keyword
import regex


def is_py3_identifier(ident):
    """Checks that ident is a valid Python 3 identifier according to
    https://docs.python.org/3/reference/lexical_analysis.html#identifiers
    """
    return bool(
        ID_REGEX.match(unicodedata.normalize('NFKC', ident)) and
        not PY3_KEYWORDS.contains(ident))

# See https://docs.python.org/3/reference/lexical_analysis.html#identifiers
ID_START_REGEX = (
    r'\p{Lu}\p{Ll}\p{Lt}\p{Lm}\p{Lo}\p{Nl}'
    r'_\u1885-\u1886\u2118\u212E\u309B-\u309C')
ID_CONTINUE_REGEX = ID_START_REGEX + (
    r'\p{Mn}\p{Mc}\p{Nd}\p{Pc}'
    r'\u00B7\u0387\u1369-\u1371\u19DA')
ID_REGEX = regex.compile(
    "[%s][%s]*$" % (ID_START_REGEX, ID_CONTINUE_REGEX), regex.UNICODE)


PY3_KEYWORDS = frozenset('False', 'None', 'True']).union(keyword.kwlist)

Note: this uses the regex package, not the built-in re package for matching against unicode categories. Also: this will reject nonlocal which is a keyword in Python 2 but not Python 3.