9 single: lexical analysis
13 A Python program is read by a *parser*. Input to the parser is a stream of
14 *tokens*, generated by the *lexical analyzer*. This chapter describes how the
15 lexical analyzer breaks a file into tokens.
17 Python uses the 7-bit ASCII character set for program text.
20 An encoding declaration can be used to indicate that string literals and
21 comments use an encoding different from ASCII.
23 For compatibility with older versions, Python only warns if it finds 8-bit
24 characters; those warnings should be corrected by either declaring an explicit
25 encoding, or using escape sequences if those bytes are binary data, instead of
28 The run-time character set depends on the I/O devices connected to the program
29 but is generally a superset of ASCII.
31 **Future compatibility note:** It may be tempting to assume that the character
32 set for 8-bit characters is ISO Latin-1 (an ASCII superset that covers most
33 western languages that use the Latin alphabet), but it is possible that in the
34 future Unicode text editors will become common. These generally use the UTF-8
35 encoding, which is also an ASCII superset, but with very different use for the
36 characters with ordinals 128-255. While there is no consensus on this subject
37 yet, it is unwise to assume either Latin-1 or UTF-8, even though the current
38 implementation appears to favor Latin-1. This applies both to the source
39 character set and the run-time character set.
47 .. index:: single: line structure
49 A Python program is divided into a number of *logical lines*.
63 The end of a logical line is represented by the token NEWLINE. Statements
64 cannot cross logical line boundaries except where NEWLINE is allowed by the
65 syntax (e.g., between statements in compound statements). A logical line is
66 constructed from one or more *physical lines* by following the explicit or
67 implicit *line joining* rules.
75 A physical line is a sequence of characters terminated by an end-of-line
76 sequence. In source files, any of the standard platform line termination
77 sequences can be used - the Unix form using ASCII LF (linefeed), the Windows
78 form using the ASCII sequence CR LF (return followed by linefeed), or the
79 Macintosh form using the ASCII CR (return) character. All of these forms can be
80 used equally, regardless of platform.
82 When embedding Python, source code strings should be passed to Python APIs using
83 the standard C conventions for newline characters (the ``\n`` character,
84 representing ASCII LF, is the line terminator).
94 single: hash character
96 A comment starts with a hash character (``#``) that is not part of a string
97 literal, and ends at the end of the physical line. A comment signifies the end
98 of the logical line unless the implicit line joining rules are invoked. Comments
99 are ignored by the syntax; they are not tokens.
104 Encoding declarations
105 ---------------------
108 single: source character set
111 If a comment in the first or second line of the Python script matches the
112 regular expression ``coding[=:]\s*([-\w.]+)``, this comment is processed as an
113 encoding declaration; the first group of this expression names the encoding of
114 the source code file. The recommended forms of this expression are ::
116 # -*- coding: <encoding-name> -*-
118 which is recognized also by GNU Emacs, and ::
120 # vim:fileencoding=<encoding-name>
122 which is recognized by Bram Moolenaar's VIM. In addition, if the first bytes of
123 the file are the UTF-8 byte-order mark (``'\xef\xbb\xbf'``), the declared file
124 encoding is UTF-8 (this is supported, among others, by Microsoft's
127 If an encoding is declared, the encoding name must be recognized by Python. The
128 encoding is used for all lexical analysis, in particular to find the end of a
129 string, and to interpret the contents of Unicode literals. String literals are
130 converted to Unicode for syntactical analysis, then converted back to their
131 original encoding before interpretation starts. The encoding declaration must
132 appear on a line of its own.
134 .. XXX there should be a list of supported encodings.
137 .. _explicit-joining:
139 Explicit line joining
140 ---------------------
143 single: physical line
145 single: line continuation
146 single: backslash character
148 Two or more physical lines may be joined into logical lines using backslash
149 characters (``\``), as follows: when a physical line ends in a backslash that is
150 not part of a string literal or comment, it is joined with the following forming
151 a single logical line, deleting the backslash and the following end-of-line
152 character. For example::
154 if 1900 < year < 2100 and 1 <= month <= 12 \
155 and 1 <= day <= 31 and 0 <= hour < 24 \
156 and 0 <= minute < 60 and 0 <= second < 60: # Looks like a valid date
159 A line ending in a backslash cannot carry a comment. A backslash does not
160 continue a comment. A backslash does not continue a token except for string
161 literals (i.e., tokens other than string literals cannot be split across
162 physical lines using a backslash). A backslash is illegal elsewhere on a line
163 outside a string literal.
166 .. _implicit-joining:
168 Implicit line joining
169 ---------------------
171 Expressions in parentheses, square brackets or curly braces can be split over
172 more than one physical line without using backslashes. For example::
174 month_names = ['Januari', 'Februari', 'Maart', # These are the
175 'April', 'Mei', 'Juni', # Dutch names
176 'Juli', 'Augustus', 'September', # for the months
177 'Oktober', 'November', 'December'] # of the year
179 Implicitly continued lines can carry comments. The indentation of the
180 continuation lines is not important. Blank continuation lines are allowed.
181 There is no NEWLINE token between implicit continuation lines. Implicitly
182 continued lines can also occur within triple-quoted strings (see below); in that
183 case they cannot carry comments.
191 .. index:: single: blank line
193 A logical line that contains only spaces, tabs, formfeeds and possibly a
194 comment, is ignored (i.e., no NEWLINE token is generated). During interactive
195 input of statements, handling of a blank line may differ depending on the
196 implementation of the read-eval-print loop. In the standard implementation, an
197 entirely blank logical line (i.e. one containing not even whitespace or a
198 comment) terminates a multi-line statement.
209 single: leading whitespace
213 single: statement grouping
215 Leading whitespace (spaces and tabs) at the beginning of a logical line is used
216 to compute the indentation level of the line, which in turn is used to determine
217 the grouping of statements.
219 First, tabs are replaced (from left to right) by one to eight spaces such that
220 the total number of characters up to and including the replacement is a multiple
221 of eight (this is intended to be the same rule as used by Unix). The total
222 number of spaces preceding the first non-blank character then determines the
223 line's indentation. Indentation cannot be split over multiple physical lines
224 using backslashes; the whitespace up to the first backslash determines the
227 **Cross-platform compatibility note:** because of the nature of text editors on
228 non-UNIX platforms, it is unwise to use a mixture of spaces and tabs for the
229 indentation in a single source file. It should also be noted that different
230 platforms may explicitly limit the maximum indentation level.
232 A formfeed character may be present at the start of the line; it will be ignored
233 for the indentation calculations above. Formfeed characters occurring elsewhere
234 in the leading whitespace have an undefined effect (for instance, they may reset
235 the space count to zero).
241 The indentation levels of consecutive lines are used to generate INDENT and
242 DEDENT tokens, using a stack, as follows.
244 Before the first line of the file is read, a single zero is pushed on the stack;
245 this will never be popped off again. The numbers pushed on the stack will
246 always be strictly increasing from bottom to top. At the beginning of each
247 logical line, the line's indentation level is compared to the top of the stack.
248 If it is equal, nothing happens. If it is larger, it is pushed on the stack, and
249 one INDENT token is generated. If it is smaller, it *must* be one of the
250 numbers occurring on the stack; all numbers on the stack that are larger are
251 popped off, and for each number popped off a DEDENT token is generated. At the
252 end of the file, a DEDENT token is generated for each number remaining on the
253 stack that is larger than zero.
255 Here is an example of a correctly (though confusingly) indented piece of Python
259 # Compute the list of all permutations of l
263 for i in range(len(l)):
267 r.append(l[i:i+1] + x)
270 The following example shows various indentation errors::
272 def perm(l): # error: first line indented
273 for i in range(len(l)): # error: not indented
275 p = perm(l[:i] + l[i+1:]) # error: unexpected indent
277 r.append(l[i:i+1] + x)
278 return r # error: inconsistent dedent
280 (Actually, the first three errors are detected by the parser; only the last
281 error is found by the lexical analyzer --- the indentation of ``return r`` does
282 not match a level popped off the stack.)
287 Whitespace between tokens
288 -------------------------
290 Except at the beginning of a logical line or in string literals, the whitespace
291 characters space, tab and formfeed can be used interchangeably to separate
292 tokens. Whitespace is needed between two tokens only if their concatenation
293 could otherwise be interpreted as a different token (e.g., ab is one token, but
302 Besides NEWLINE, INDENT and DEDENT, the following categories of tokens exist:
303 *identifiers*, *keywords*, *literals*, *operators*, and *delimiters*. Whitespace
304 characters (other than line terminators, discussed earlier) are not tokens, but
305 serve to delimit tokens. Where ambiguity exists, a token comprises the longest
306 possible string that forms a legal token, when read from left to right.
311 Identifiers and keywords
312 ========================
318 Identifiers (also referred to as *names*) are described by the following lexical
322 identifier: (`letter`|"_") (`letter` | `digit` | "_")*
323 letter: `lowercase` | `uppercase`
328 Identifiers are unlimited in length. Case is significant.
338 single: reserved word
340 The following identifiers are used as reserved words, or *keywords* of the
341 language, and cannot be used as ordinary identifiers. They must be spelled
342 exactly as written here::
344 and del from not while
345 as elif global or with
346 assert else if pass yield
347 break except import print
349 continue finally is return
352 .. versionchanged:: 2.4
353 :const:`None` became a constant and is now recognized by the compiler as a name
354 for the built-in object :const:`None`. Although it is not a keyword, you cannot
355 assign a different object to it.
357 .. versionchanged:: 2.5
358 Both :keyword:`as` and :keyword:`with` are only recognized when the
359 ``with_statement`` future feature has been enabled. It will always be enabled in
360 Python 2.6. See section :ref:`with` for details. Note that using :keyword:`as`
361 and :keyword:`with` as identifiers will always issue a warning, even when the
362 ``with_statement`` future directive is not in effect.
367 Reserved classes of identifiers
368 -------------------------------
370 Certain classes of identifiers (besides keywords) have special meanings. These
371 classes are identified by the patterns of leading and trailing underscore
375 Not imported by ``from module import *``. The special identifier ``_`` is used
376 in the interactive interpreter to store the result of the last evaluation; it is
377 stored in the :mod:`__builtin__` module. When not in interactive mode, ``_``
378 has no special meaning and is not defined. See section :ref:`import`.
382 The name ``_`` is often used in conjunction with internationalization;
383 refer to the documentation for the :mod:`gettext` module for more
384 information on this convention.
387 System-defined names. These names are defined by the interpreter and its
388 implementation (including the standard library); applications should not expect
389 to define additional names using this convention. The set of names of this
390 class defined by Python may be extended in future versions. See section
394 Class-private names. Names in this category, when used within the context of a
395 class definition, are re-written to use a mangled form to help avoid name
396 clashes between "private" attributes of base and derived classes. See section
397 :ref:`atom-identifiers`.
409 Literals are notations for constant values of some built-in types.
417 .. index:: single: string literal
419 String literals are described by the following lexical definitions:
421 .. index:: single: ASCII@ASCII
424 stringliteral: [`stringprefix`](`shortstring` | `longstring`)
425 stringprefix: "r" | "u" | "ur" | "R" | "U" | "UR" | "Ur" | "uR"
426 shortstring: "'" `shortstringitem`* "'" | '"' `shortstringitem`* '"'
427 longstring: "'''" `longstringitem`* "'''"
428 : | '"""' `longstringitem`* '"""'
429 shortstringitem: `shortstringchar` | `escapeseq`
430 longstringitem: `longstringchar` | `escapeseq`
431 shortstringchar: <any source character except "\" or newline or the quote>
432 longstringchar: <any source character except "\">
433 escapeseq: "\" <any ASCII character>
435 One syntactic restriction not indicated by these productions is that whitespace
436 is not allowed between the :token:`stringprefix` and the rest of the string
437 literal. The source character set is defined by the encoding declaration; it is
438 ASCII if no encoding declaration is given in the source file; see section
442 single: triple-quoted string
443 single: Unicode Consortium
444 single: string; Unicode
447 In plain English: String literals can be enclosed in matching single quotes
448 (``'``) or double quotes (``"``). They can also be enclosed in matching groups
449 of three single or double quotes (these are generally referred to as
450 *triple-quoted strings*). The backslash (``\``) character is used to escape
451 characters that otherwise have a special meaning, such as newline, backslash
452 itself, or the quote character. String literals may optionally be prefixed with
453 a letter ``'r'`` or ``'R'``; such strings are called :dfn:`raw strings` and use
454 different rules for interpreting backslash escape sequences. A prefix of
455 ``'u'`` or ``'U'`` makes the string a Unicode string. Unicode strings use the
456 Unicode character set as defined by the Unicode Consortium and ISO 10646. Some
457 additional escape sequences, described below, are available in Unicode strings.
458 The two prefix characters may be combined; in this case, ``'u'`` must appear
461 In triple-quoted strings, unescaped newlines and quotes are allowed (and are
462 retained), except that three unescaped quotes in a row terminate the string. (A
463 "quote" is the character used to open the string, i.e. either ``'`` or ``"``.)
466 single: physical line
467 single: escape sequence
471 Unless an ``'r'`` or ``'R'`` prefix is present, escape sequences in strings are
472 interpreted according to rules similar to those used by Standard C. The
473 recognized escape sequences are:
475 +-----------------+---------------------------------+-------+
476 | Escape Sequence | Meaning | Notes |
477 +=================+=================================+=======+
478 | ``\newline`` | Ignored | |
479 +-----------------+---------------------------------+-------+
480 | ``\\`` | Backslash (``\``) | |
481 +-----------------+---------------------------------+-------+
482 | ``\'`` | Single quote (``'``) | |
483 +-----------------+---------------------------------+-------+
484 | ``\"`` | Double quote (``"``) | |
485 +-----------------+---------------------------------+-------+
486 | ``\a`` | ASCII Bell (BEL) | |
487 +-----------------+---------------------------------+-------+
488 | ``\b`` | ASCII Backspace (BS) | |
489 +-----------------+---------------------------------+-------+
490 | ``\f`` | ASCII Formfeed (FF) | |
491 +-----------------+---------------------------------+-------+
492 | ``\n`` | ASCII Linefeed (LF) | |
493 +-----------------+---------------------------------+-------+
494 | ``\N{name}`` | Character named *name* in the | |
495 | | Unicode database (Unicode only) | |
496 +-----------------+---------------------------------+-------+
497 | ``\r`` | ASCII Carriage Return (CR) | |
498 +-----------------+---------------------------------+-------+
499 | ``\t`` | ASCII Horizontal Tab (TAB) | |
500 +-----------------+---------------------------------+-------+
501 | ``\uxxxx`` | Character with 16-bit hex value | \(1) |
502 | | *xxxx* (Unicode only) | |
503 +-----------------+---------------------------------+-------+
504 | ``\Uxxxxxxxx`` | Character with 32-bit hex value | \(2) |
505 | | *xxxxxxxx* (Unicode only) | |
506 +-----------------+---------------------------------+-------+
507 | ``\v`` | ASCII Vertical Tab (VT) | |
508 +-----------------+---------------------------------+-------+
509 | ``\ooo`` | Character with octal value | (3,5) |
511 +-----------------+---------------------------------+-------+
512 | ``\xhh`` | Character with hex value *hh* | (4,5) |
513 +-----------------+---------------------------------+-------+
515 .. index:: single: ASCII@ASCII
520 Individual code units which form parts of a surrogate pair can be encoded using
521 this escape sequence.
524 Any Unicode character can be encoded this way, but characters outside the Basic
525 Multilingual Plane (BMP) will be encoded using a surrogate pair if Python is
526 compiled to use 16-bit code units (the default). Individual code units which
527 form parts of a surrogate pair can be encoded using this escape sequence.
530 As in Standard C, up to three octal digits are accepted.
533 Unlike in Standard C, exactly two hex digits are required.
536 In a string literal, hexadecimal and octal escapes denote the byte with the
537 given value; it is not necessary that the byte encodes a character in the source
538 character set. In a Unicode literal, these escapes denote a Unicode character
539 with the given value.
541 .. index:: single: unrecognized escape sequence
543 Unlike Standard C, all unrecognized escape sequences are left in the string
544 unchanged, i.e., *the backslash is left in the string*. (This behavior is
545 useful when debugging: if an escape sequence is mistyped, the resulting output
546 is more easily recognized as broken.) It is also important to note that the
547 escape sequences marked as "(Unicode only)" in the table above fall into the
548 category of unrecognized escapes for non-Unicode string literals.
550 When an ``'r'`` or ``'R'`` prefix is present, a character following a backslash
551 is included in the string without change, and *all backslashes are left in the
552 string*. For example, the string literal ``r"\n"`` consists of two characters:
553 a backslash and a lowercase ``'n'``. String quotes can be escaped with a
554 backslash, but the backslash remains in the string; for example, ``r"\""`` is a
555 valid string literal consisting of two characters: a backslash and a double
556 quote; ``r"\"`` is not a valid string literal (even a raw string cannot end in
557 an odd number of backslashes). Specifically, *a raw string cannot end in a
558 single backslash* (since the backslash would escape the following quote
559 character). Note also that a single backslash followed by a newline is
560 interpreted as those two characters as part of the string, *not* as a line
563 When an ``'r'`` or ``'R'`` prefix is used in conjunction with a ``'u'`` or
564 ``'U'`` prefix, then the ``\uXXXX`` and ``\UXXXXXXXX`` escape sequences are
565 processed while *all other backslashes are left in the string*. For example,
566 the string literal ``ur"\u0062\n"`` consists of three Unicode characters: 'LATIN
567 SMALL LETTER B', 'REVERSE SOLIDUS', and 'LATIN SMALL LETTER N'. Backslashes can
568 be escaped with a preceding backslash; however, both remain in the string. As a
569 result, ``\uXXXX`` escape sequences are only recognized when there are an odd
570 number of backslashes.
573 .. _string-catenation:
575 String literal concatenation
576 ----------------------------
578 Multiple adjacent string literals (delimited by whitespace), possibly using
579 different quoting conventions, are allowed, and their meaning is the same as
580 their concatenation. Thus, ``"hello" 'world'`` is equivalent to
581 ``"helloworld"``. This feature can be used to reduce the number of backslashes
582 needed, to split long strings conveniently across long lines, or even to add
583 comments to parts of strings, for example::
585 re.compile("[A-Za-z_]" # letter or underscore
586 "[A-Za-z0-9_]*" # letter, digit or underscore
589 Note that this feature is defined at the syntactical level, but implemented at
590 compile time. The '+' operator must be used to concatenate string expressions
591 at run time. Also note that literal concatenation can use different quoting
592 styles for each component (even mixing raw strings and triple quoted strings).
602 single: numeric literal
603 single: integer literal
604 single: plain integer literal
605 single: long integer literal
606 single: floating point literal
607 single: hexadecimal literal
608 single: octal literal
609 single: decimal literal
610 single: imaginary literal
611 single: complex; literal
613 There are four types of numeric literals: plain integers, long integers,
614 floating point numbers, and imaginary numbers. There are no complex literals
615 (complex numbers can be formed by adding a real number and an imaginary number).
617 Note that numeric literals do not include a sign; a phrase like ``-1`` is
618 actually an expression composed of the unary operator '``-``' and the literal
624 Integer and long integer literals
625 ---------------------------------
627 Integer and long integer literals are described by the following lexical
631 longinteger: `integer` ("l" | "L")
632 integer: `decimalinteger` | `octinteger` | `hexinteger`
633 decimalinteger: `nonzerodigit` `digit`* | "0"
634 octinteger: "0" `octdigit`+
635 hexinteger: "0" ("x" | "X") `hexdigit`+
636 nonzerodigit: "1"..."9"
638 hexdigit: `digit` | "a"..."f" | "A"..."F"
640 Although both lower case ``'l'`` and upper case ``'L'`` are allowed as suffix
641 for long integers, it is strongly recommended to always use ``'L'``, since the
642 letter ``'l'`` looks too much like the digit ``'1'``.
644 Plain integer literals that are above the largest representable plain integer
645 (e.g., 2147483647 when using 32-bit arithmetic) are accepted as if they were
646 long integers instead. [#]_ There is no limit for long integer literals apart
647 from what can be stored in available memory.
649 Some examples of plain integer literals (first row) and long integer literals
650 (second and third rows)::
653 3L 79228162514264337593543950336L 0377L 0x100000000L
654 79228162514264337593543950336 0xdeadbeef
659 Floating point literals
660 -----------------------
662 Floating point literals are described by the following lexical definitions:
665 floatnumber: `pointfloat` | `exponentfloat`
666 pointfloat: [`intpart`] `fraction` | `intpart` "."
667 exponentfloat: (`intpart` | `pointfloat`) `exponent`
669 fraction: "." `digit`+
670 exponent: ("e" | "E") ["+" | "-"] `digit`+
672 Note that the integer and exponent parts of floating point numbers can look like
673 octal integers, but are interpreted using radix 10. For example, ``077e010`` is
674 legal, and denotes the same number as ``77e10``. The allowed range of floating
675 point literals is implementation-dependent. Some examples of floating point
678 3.14 10. .001 1e100 3.14e-10 0e0
680 Note that numeric literals do not include a sign; a phrase like ``-1`` is
681 actually an expression composed of the unary operator ``-`` and the literal
690 Imaginary literals are described by the following lexical definitions:
693 imagnumber: (`floatnumber` | `intpart`) ("j" | "J")
695 An imaginary literal yields a complex number with a real part of 0.0. Complex
696 numbers are represented as a pair of floating point numbers and have the same
697 restrictions on their range. To create a complex number with a nonzero real
698 part, add a floating point number to it, e.g., ``(3+4j)``. Some examples of
701 3.14j 10.j 10j .001j 1e100j 3.14e-10j
709 .. index:: single: operators
711 The following tokens are operators::
717 The comparison operators ``<>`` and ``!=`` are alternate spellings of the same
718 operator. ``!=`` is the preferred spelling; ``<>`` is obsolescent.
726 .. index:: single: delimiters
728 The following tokens serve as delimiters in the grammar::
735 The period can also occur in floating-point and imaginary literals. A sequence
736 of three periods has a special meaning as an ellipsis in slices. The second half
737 of the list, the augmented assignment operators, serve lexically as delimiters,
738 but also perform an operation.
740 The following printing ASCII characters have special meaning as part of other
741 tokens or are otherwise significant to the lexical analyzer::
745 .. index:: single: ASCII@ASCII
747 The following printing ASCII characters are not used in Python. Their
748 occurrence outside string literals and comments is an unconditional error::
752 .. rubric:: Footnotes
754 .. [#] In versions of Python prior to 2.4, octal and hexadecimal literals in the range
755 just above the largest representable plain integer but below the largest
756 unsigned 32-bit number (on a machine using 32-bit arithmetic), 4294967296, were
757 taken as the negative plain integer obtained by subtracting 4294967296 from
758 their unsigned value.