<$marker2><00> (Continued)X<00>+ (Sheet <$tblsheetnum> of <$tblsheetcount>) te<00>Heading & Page<20><$paratext><3E> on page<$pagenum><00>Pagepage<$pagenum>Z<00>See Heading & Page%See <20><$paratext><3E> on page<$pagenum>.px <00> Table All\7Table<$paranumonly>, <20><$paratext>,<2C> on page<$pagenum>r<00>Table Number & Pagea'Table<$paranumonly> on page<$pagenum>S<03><03><03> 40<03><0E><0E><03><03>Aer<03>Aapa<03>4Adnu<03>==<03>??<03>AAAe> <03>CC<03>u<03><00><03>/um
<EFBFBD>/USUT<03><>(<03>k<00>A Python program is read by a parserk. Input to the parser is a stream of tokensk, generated by the lex<UQUT<03><>@<03><0B>^ical analyzerk. This chapter describes how the lexical analyzer breaks a file into tokens.
UUOUT<03><><03>khPython uses the 7-bit ASCII character set for program text and string literals. 8-bit characters may be
bUMUT<03><><03>k<>gused in string literals and comments but their interpretation is platform dependent; the proper way to oUKUT<03><>@<03>k^insert 8-bit characters in string literals is by using octal or hexadecimal escape sequences.
<00>UIUT<03><><03>keThe run-time character set depends on the I/O devices connected to the program but is generally a su<00>UGUT<03><>@<03>kperset of ASCII.
<00>UEUT<03><><03>ehFuture compatibility note: kIt may be tempting to assume that the character set for 8-bit characters <00>UCUT<03><><03>kcis ISO Latin-1 (an ASCII superset that covers most western languages that use the Latin alphabet), n p<00>UAUT<03><><03>ka gbut it is possible that in the future Unicode text editors will become common. These generally use the UT<00>U?UT<03><><03>k aeUTF-8 encoding, which is also an ASCII superset, but with very different use for the characters with <20><00>U=UT<03><><03>kthgordinals 128-255. While there is no consensus on this subject yet, it is unwise to assume either Latin-<03><00>U;UT<03><><03>kngf1 or UTF-8, even though the current implementation appears to favor Latin-1. This applies both to the <00>U9UT<03><>@<03>k-b5source character set and the run-time character set.
eU7 UT<04><>`<03>ieqLine structure
<03><>0U5UT<03><>h<03>ktiRA Python program is divided in a number of logical linesk.k
I<><49>`<03>l<><6C>Logical lines
]U3UT<03><>(<03>kUTjThe end of each logical line is represented by the token kNEWLINE. Statements cannot cross logical jU1UT<03><><03>kUT^line boundaries except where NEWLINE is allowed by the syntax (e.g. between statements in comwU/UT<03><><03>k pupound statements). A logical line is constructed from one or more physical linesk by following the exe<01>U-UT<03><>
@<03>k<><6B>3plicit or implicit line joiningk rules.
rse<01><><EFBFBD>
`<03>lifPhysical lines
e c<01>U+UT<03><>
<03>kU=fA physical line ends in whatever the current platform<72>s convention is for terminating lines. On UNIX, <01>U)UT<03><><03>k<03>bthis is the ASCII LF (linefeed) character. On DOS/Windows, it is the ASCII sequence CR LF (return <01>U'UT<03><>@<03>ko Lfollowed by linefeed). On Macintosh, it is the ASCII CR (return) character.
(<03>k<03>xA comment starts with a khash character (<00>#k) that is not part of a string literal, and ends at the end of s
U#UT<03><>
<03>kUTjthe physical line. A comment signifies the end of the logical line unless the implicit line joining rules U!UT<03><>
@<03>kUTHare invoked. Comments are ignored by the syntax<61><78>; they are not tokens.
ee+<2B><>
h<03>lmExplicit lline joining
?UUT<03><>
(<03>k lxTwo or more physical lines may be joined into logical lines using k$backslash characters (<00>\k), as follows: plLUUT<03><><03>kinlwhen a physical line ends in a backslash that is not part of a string literal or comment, it is joined with evYUUT<03><><03>ktfwthe following forming a single k&logical line, deleting the backslash and the followingk' end-of-line charOS/fUUT<03><>@<03>k Aacter. For example:
etx<>m<06><>UU`<03>s<03>.if 1900 < year < 2100 and 1 <= month <= 12 \
<02><>l<06><>UU`<03>sur, and 1 <= day <= 31 and 0 <= hour < 24 \
@<03>kUT" C&<06><><06><>UU`<03>s sK and 0 <= minute < 60 and 0 <= second < 60: # Looks like a valid date
oin<12><><06><>UU`<03>s<03> return 1
+URUT<03><><03>ke eA line ending in a backslash cannot carry a comment. A backslash does not continue a comment. A back<06>8UPUT<03><><03>kcatslash does not continue a token except for string literals (i.e., tokens other than string literals cannot be split thEUNUT<03><>H<03>kg wacross physical k(lines using a backslash). A backslash is illegal elsewhere on a line outside a string literal.
ac^<5E><>h<03>letImplicit l)line joining
rULUT<03><>(<03>k<=zExpressions in parentheses, k+square brackets or k*curly braces can be split over more than one physical line UJUT<03><>@<03>k(without using backslashes. For example:
<00><><EFBFBD><06><>UU`<03>sEmonth_names = [<5B>Januari<72>, <20>Februari<72>, <20>Maart<72>, # These are the
&<00><><EFBFBD><06><>UU`<03>sE <20>Juli<6C>, <20>Augustus<75>, <20>September<65>, # for the months
n<00><><EFBFBD><06><>UU`<03>sliC <20>Oktober<65>, <20>November<65>, <20>December<65>] # of the year
<03><><00>UDUT<03><>(<03>kndsImplicitly continued lines can carry comments. The k,indentation of the continuation lines is not important. sh <00>UBUT<03><><03>ka pBlank continuation lines are allowed. There is no k-NEWLINE token between implicit continuation lines. Imac<00>U@UT<03><><03>kwplicit continued lines can also occur within k.triple-quoted strings (see below); in that case they cannot carry Im<00>U>UT<03><>@<03>ke
th"U<UT<03><><03>krelA logical line that contains only spaces, tabs, formfeeds, and possibly a comment, is ignored (i.e., no NEWus/U:UT<03><><03>kormLINE token is generated), except that during interactive input of statements, an entirely blank logical line <20><U8UT<03><>@<03>k Z(i.e. one containing not even whitespace or a comment) terminates a multi-line statement.
U<><55>h<03>l 0Indentation
mbiU6UT<03><>(<03>knt<00>3Leading k4whitespace (k6spaces and k7tabs) at the beginning of a logical line is used to compute the indentation d lvU4UT<03><>H<03>kmeVlevel of the line, which in turn is used to determine the grouping of 5statements.
<01>U2UT<03><><03>kntqFirst, tabs are replaced (from left to right) by one to eight spaces such that the total number of characters up <20><01>U0UT<03><><03>konpto there is a multiple of eight (this is intended to be the same rule as used by 8UNIX). The total number of U><01>U.UT<03><><03>kcomspaces preceding the first non-blank character then determines the line<6E>s indentation. Indentation cannot be <01>U,UT<03><><03>kndnsplit over multiple physical lines using backslashes; the whitespace up to the first backslash determines the <01>U*UT<03><>@<03>kut
<01>U(UT<03><><03>elomCross-platform compatibility note:k because of the nature of text editors on non-UNIX platforms, it is une<01>U&UT<03><>@<03>kVwise to use a mixture of spaces and tabs for the indentation in a single source file.
U$UT<03><>(<03>kkuA k9formfeed character may be present at the start of the line; formfeed characters occurring elsewhere in the lU"UT<03><>H<03>k<19>lleading whitespace have an undefined effect (for instance, they may reset the k;space count to zero).
to(U UT<03><>(<03>keikThe indentation levels of consecutive lines are used to generate k2INDENT and DEDENT tokens, using a (t5UUT<03><>@<03>k bstack, as follows.
d NUUT<03><>(<03>ktowBefore the first line of the file is read, a single zero is pushed on the k<stack; this will never be popped off on.[UUT<03><><03>kt kagain. The numbers pushed on the stack will always be strictly increasing from bottom to top. At the begint bhUUT<03><><03>ks rning of each logical line, the line<6E>s indentation level is compared to the top of the stack. If it is equal, nothuUUT<03><><03>ktoqing happens. If it is larger, it is pushed on the stack, and one INDENT token is generated. If it is smaller, it iHH<01><02><01>"><13>kHH<01><02>l<0E><>%!!<03> thd<01>#=he %%{<19>lecH<01><02><00>$>#f icH<01><02><02><><EFBFBD>t **to zero)%<00><>*UTUT<03><><03>lekmustk be one of the numbers occurring on the stack; all numbers on the stack that are larger are popped k bURUT<03><><03>ks.boff, and for each number popped off a DEDENT token is generated. At the end of the file, a DEDENT !UPUT<03><>@<03>ks Utoken is generated for each number remaining on the stack that is larger than zero.
k4UNUT<03><>`<03>kriVHere is an example of a correctly (though confusingly) indented piece of Python code:
F<><46><06><>UU`<03>sindef perm(l):
R<><52><06><>UU`<03>sto5 # Compute the list of all permutations of l
<EFBFBD>^<5E><><06><>UU`<03>sen if len(l) <= 1:
ij<><6A><06><>UU`<03>sac return [l]
v<><76><06><>UU`<03>sll r = []
<00><><EFBFBD><06><>UU`<03>s> for i in range(len(l)):
<00>UAUT<03><>h<03>kUT@The following example shows various k?indentation errors:
k;骕<06><>UU`<03>se F def perm(l): # error: first line indented
<00><><EFBFBD><06><>UU`<03>sof? for i in range(len(l)): # error: not indented
!UP<01><><06><>UU`<03>sto s = l[:i] + l[i+1:]
r
<EFBFBD><EFBFBD><06><>UU`<03>sstD p = perm(l[:i] + l[i+1:]) # error: unexpected indent
le<19><><06><>UU`<03>sho for x in p:
i%<25><><06><>UU`<03>s:
, r.append(l[i:i+1] + x)
UU1<><31><06><>UU`<03>s# F return r # error: inconsistent dedent
JU8UT<03><>(<03>k<><6B>}(Actually, the first three errors are detected by the k>parser; only the last error is found by the k=lexical annWU6UT<03><>@<03>kUU[alyzer <20> the indentation of <00>return rk does not match a level popped off the stack.)
<EFBFBD><EFBFBD><EFBFBD>p<><70>`<03>l<19>Whitespace between tokens
<01>U4UT<03><><03>k jExcept at the beginning of a logical line or in string literals, the whitespace characters space, tab and <01>U2UT<03><><03>kouaformfeed can be used interchangeably to separate tokens. Whitespace is needed between two tokens o<01>U0UT<03><><03>kntponly if their concatenation could otherwise be interpreted as a different token (e.g., <00>abk is one token, <01>U.UT<03><>@<03>kl[(but <00>ak<00>bk is two tokens).
u<01>U*UT<03><>(<03>klekBesides NEWLINE, INDENT and DEDENT, the following categories of tokens exist: k@identifiersk, x)<01>U(UT<03><><03><03><00>keywordsk, literalsk, operatorsk, and delimitersk. Whitespace characters (other than line terminators, disre <01>U&UT<03><><03>kkgcussed earlier) are not tokens, but serve to delimit tokens. Where ambiguity exists, a token comprises denU$UT<03><>@<03>krnSthe longest possible string that forms a legal token when read from left to right.
ace'U" UT<04><>h<03>iU4BIdentifiers and keywords
@U UT<03><>h<03>kichIdentifiers (also referred to as Anamesk) are described by the following lexical definitions:
UTcH<01><02><01>%@#<13>lecH<01><02>l<0E><>"($$<03>s od<01>&=,x)((|orHH<01><02><00>'>&dHH<01><02>y<><79>span line 7<0E>(isUTUT<03><>`<03>kcu:Identifiers are unlimited in length. Case is significant.
<20><>`<03>lst Keywords
c4URUT<03><>(<03>kUT}The following identifiers are used as kDreserved words, or Ckeywordsk of the language, and cannot be used as dAUPUT<03><>@<03>korDordinary identifiers. They must be spelled exactly as written here:
esS<><53><06><>UU`<03>sd 2and del for is raise
_<><5F><06><>UU`<03>ser2assert elif from lambda return
sy Kshortstringchar: <any ASCII character except "\" or newline or the quote>
nti<1D><><06><>UU`s3longstringchar: <any ASCII character except "\">
ide)<29><><06><>UUhspe4escapeseq: "\" <any ASCII character> sN
BU7UT<03><>(
kn rIn plain English: String literals can be enclosed in single Mquotes (<28>) or double quotes ("). They can also be OU5UT<03><>
kf |enclosed in groups of three single or double quotes (these are generally referred to as Ltriple-quoted stringsk). lo\U3UT<03><>
kitpThe Kbackslash (\) character is used to escape characters that otherwise have a special meaning, such as newreiU1UT<03><>
kR"pline, backslash itself, or the quote character. String literals may optionally be prefixed with a letter <20>r<EFBFBD> or vU/UT<03><>@
k "i<>R<EFBFBD>; such strings are called raw stringsk and use different rules for backslash escape sequences.
iHH<01><02><01>(>&ostHH<01><02>l<0E><>%+''<03>Kshd<01>)=" o++}ti<00><>cH<01><02><00>*>)IchcH<01><02>F<><46>decapeseq:5<0E>+y UTUT<03><>keIn <20>long strings<67> (strings surrounded by sets of three quotes), unescaped newlines and quotes are albURUT<03><>ky glowed (and are retained), except that three unescaped quotes in a row terminate the string. (A <20>quote<74> y r!UPUT<03><>@kOis the character used to open the string, i.e. either <00><1B>k or <00>"k.)
cte:UNUT<03><>(ke kOUnless an <20>r<EFBFBD> or <20>R<EFBFBD> prefix is present, escape sequences in strings are interpreted according to rules orGULUT<03><>@HkerRsimilar to those used by Standard QC. The recognized escape sequences are::
<01>UJUT<03><>krihIn strict compatibility with Standard C, up to three octal digits are accepted, but an unlimited number <01>UHUT<03><>khof hex digits is taken to be part of the hex escape (and then the lower 8 bits of the resulting hex num<01>UFUT<03><>@k2ber are used in all current implementations...).
<01>UDUT<03><>(kInUnlike Standard C, all Sunrecognized escape sequences are left in the string unchanged, i.e., the back<01>UBUT<03><>gsmslash is left in the string.k (This behavior is useful when debugging: if an escape sequence is mistyped, eU@UT<03><>@k t;the resulting output is more easily recognized as broken.)
<EFBFBD>quU>UT<03><><03>k<><6B>rWhen an <20>r<EFBFBD> or <20>R<EFBFBD> prefix is present, all backslashes are left in the string.k For example, the string lit)U<UT<03><><03>k <20>meral <00>r"\n"k consists of two characters: a backslash and a lowercase <20>n<EFBFBD>. String quotes can be escaped i6U:UT<03><><03>kd pwith a backslash, but the backslash remains in the string; for example, <00>r"\""k is a valid string literal ndCU8UT<03><>@<03>k o>consisting of two characters: a backslash and a double quote.
URUT<03><>(kismMultiple adjacent kTstring literals (delimited by whitespace), possibly using different quoting convene-UPUT<03><>ks ptions, are allowed, and their meaning is the same as their concatenation. Thus, <00>"hello"k<00><1B>world<6C>k .:UNUT<03><>khedis equivalent to <00>"helloworld"k. This feature can be used to reduce the number of backslashes erGULUT<03><>kuobneeded, to split long strings conveniently across long lines, or even to add comments to parts of TUJUT<03><>@kr"strings, for example:
f<><66><06><>UU`sUT7regex.compile("[A-Za-z_]" # letter or underscore
nd r<><72><06><>UU`s> "[A-Za-z0-9_]*" # letter, digit or underscore
~<7E><><06><>UU`s )
<EFBFBD><00>UEUT<03><>khNote that this feature is defined at the syntactical level, but implemented at compile time. The <20>+<2B> opS<00>UCUT<03><>kherator must be used to concatenate string expressions at run time. Also note that literal concatenation te<00>UAUT<03><>@k<><6B>5can use different quoting styles for each component.
l<00><><EFBFBD>hlceUNumeric literals
r<00>U?UT<03><>keeThere are four types of numeric literals: plain integers, long integers, floating point numbers, and <00>U=UT<03><>Hk<>wimaginary kVnumbers.
kheU;UT<03><>hn .WInteger andnX long integer literals
U9UT<03><>hk o[Integer and long integer literals are described by the following lexical definitions: Y
n t<01>U/UT<03><>$kt hAlthough both lower case <20>l<EFBFBD> and upper case <20>L<EFBFBD> are allowed as suffix for long integers, it is strongly <01>U-UT<03><>@$kNuXrecommended to always use <20>L<EFBFBD>, since the letter <20>l<EFBFBD> looks too much like the digit <20>1<EFBFBD>.
ge<01>U+UT<03><>%k<><6B>hPlain integer decimal literals must be at most 2147483647 (i.e., the largest positive integer, using 32-In<01>U)UT<03><>%klogbit arithmetic). Plain octal and hexadecimal literals may be as large as 4294967295, but values larger g l<01>U'UT<03><>%k: cthan 2147483647 are converted to a negative value by subtracting 4294967296. There is no limit for <01>U%UT<03><>@%k oJlong integer literals apart from what can be stored in available memory.
<01>U#UT<03><>`&k<><6B>2Some examples of plain and long integer literals:
UU+URUT<03><>1k47eNote that the integer part of a floating point number cannot look like an octal integer. The allowed 98UPUT<03><>1k0xerange of floating point literals is implementation-dependent. Some examples of floating point litertEUNUT<03><>@1kd als:
<00>UFUT<03><>(6kitiAn imaginary literals yields a complex number with a real part of 0.0. k\Complex numbers are repre1<00>UDUT<03><>6kfsented as a pair of floating point numbers and have the same restrictions on their range. To create a <00>UBUT<03><>6kicomplex number with a nonzero real part, add a floating point number to it, e.g. <00>(3+4j).k Some ex<00>U@UT<03><>@6k&les of imaginary literals:
nteU=UT<03><>(8ktivNote that numeric literals do not include a k]sign; a phrase like <00>-1k is actually an expression composed U;UT<03><>@8kep=of the unary operator <20><00>-k<1B> and the literal <00>1k.
DU7UT<03><>h:k.0(The following tokens are ^operators:
eV<><56><06><>UU`;sls++ - * ** / %
alsb<><62><06><>UU`<sth+<< >> & | ^ ~
UUn<><6E><06><>UU`=ser4< > <= >= == != <>
6<01>U2UT<03><>(>k l}The k_comparison operators <00><>k and <00>!=k are alternate spellings of the same operator; != is the preferred a<01>U0UT<03><>@>kntspelling, <> is obsolescent.
<01>U. UT<04><>`?ianDelimiters
a <01>U,UT<03><>h@k=The following tokens serve as `delimiters in the grammar:
ڪ<><06><>UU`As +( ) [ ] { }
am<01><><06><>UU`Bsli+, : . <20> = ;
10<01>U(UT<03><>Ck0jfThe period can also occur in floating-point and imaginary literals. A sequence of three periods has a U&UT<03><>@Cks 'special meaning as ellipses in slices.
<03><>%U$UT<03><>DknabThe following printing ASCII characters have special meaning as part of other tokens or are other2U"UT<03><>@Dkg *wise significant to the lexical analyzer:
D<>v<06><>UU`Es <1B> " # \
als]UUT<03><>(FkthlThe following printing kaASCII characters are not used in Python. Their occurrence outside string lit!=jUUT<03><>@FkUT.erals and comments is an unconditional error:
|<7C>q<06><>UU`Gs!=@ $ ? ellcH<01><02><03>1@/r acH<01><02>l<0E><>.400<03><04><>d<03>2=Th44<0E>s `L<>H<01><02><00>3>2AL<>H<01><02><01><><06><>4, L<>H<01><02><03>4@2T<00><>L<>H<01><02>l<0E><>133<03>ginc<00><><EFBFBD>;<09>5><0E>)<00>aa c<00><><EFBFBD>;sp*:<02>ninas`<03>gs.Escape SequencesDʀ<00><><EFBFBD>;<09>6>)<00>rhaʀ<00><><EFBFBD>;pa*:<02>er kee<03>gU"TH<1A><>D<09>7><0E>&<00>aanH<1A><>DE'1<02>" `Hgls Special Meanings of Identifiersg <00><1A><>D<09>8>&<00>- i<00><1A><>Den'1<02>e sineIg<><67>Fd<09><=lrr=Au@ L<>$<01><09>=>?<@L<>$<01>l<0E><>>><03><0E><>L<>$<01><09>>><=L<>$<01>L<>H=<00>`Kp<><0B>L<><02><01><09>?>=A<<13>L<><02><01>l<0E><>@@<03>TL<><02><01><09>@><L<><02><01><09><><0E>)?<06><><06><>UUhLu<02>
