Updated for 2.1b2 distribution.
[python/dscho.git] / Doc / lib / libtime.tex
blobc3f62ed7bcbf79fa087cced388b229c42c715ffe
1 \section{\module{time} ---
2 Time access and conversions}
4 \declaremodule{builtin}{time}
5 \modulesynopsis{Time access and conversions.}
8 This module provides various time-related functions.
9 It is always available, but not all functions are available
10 on all platforms.
12 An explanation of some terminology and conventions is in order.
14 \begin{itemize}
16 \item
17 The \dfn{epoch}\index{epoch} is the point where the time starts. On
18 January 1st of that year, at 0 hours, the ``time since the epoch'' is
19 zero. For \UNIX{}, the epoch is 1970. To find out what the epoch is,
20 look at \code{gmtime(0)}.
22 \item
23 The functions in this module do not handle dates and times before the
24 epoch or far in the future. The cut-off point in the future is
25 determined by the C library; for \UNIX{}, it is typically in
26 2038\index{Year 2038}.
28 \item
29 \strong{Year 2000 (Y2K) issues}:\index{Year 2000}\index{Y2K} Python
30 depends on the platform's C library, which generally doesn't have year
31 2000 issues, since all dates and times are represented internally as
32 seconds since the epoch. Functions accepting a time tuple (see below)
33 generally require a 4-digit year. For backward compatibility, 2-digit
34 years are supported if the module variable \code{accept2dyear} is a
35 non-zero integer; this variable is initialized to \code{1} unless the
36 environment variable \envvar{PYTHONY2K} is set to a non-empty string,
37 in which case it is initialized to \code{0}. Thus, you can set
38 \envvar{PYTHONY2K} to a non-empty string in the environment to require 4-digit
39 years for all year input. When 2-digit years are accepted, they are
40 converted according to the \POSIX{} or X/Open standard: values 69-99
41 are mapped to 1969-1999, and values 0--68 are mapped to 2000--2068.
42 Values 100--1899 are always illegal. Note that this is new as of
43 Python 1.5.2(a2); earlier versions, up to Python 1.5.1 and 1.5.2a1,
44 would add 1900 to year values below 1900.
46 \item
47 UTC\index{UTC} is Coordinated Universal Time\index{Coordinated
48 Universal Time} (formerly known as Greenwich Mean
49 Time,\index{Greenwich Mean Time} or GMT). The acronym UTC is not a
50 mistake but a compromise between English and French.
52 \item
53 DST is Daylight Saving Time,\index{Daylight Saving Time} an adjustment
54 of the timezone by (usually) one hour during part of the year. DST
55 rules are magic (determined by local law) and can change from year to
56 year. The C library has a table containing the local rules (often it
57 is read from a system file for flexibility) and is the only source of
58 True Wisdom in this respect.
60 \item
61 The precision of the various real-time functions may be less than
62 suggested by the units in which their value or argument is expressed.
63 E.g.\ on most \UNIX{} systems, the clock ``ticks'' only 50 or 100 times a
64 second, and on the Mac, times are only accurate to whole seconds.
66 \item
67 On the other hand, the precision of \function{time()} and
68 \function{sleep()} is better than their \UNIX{} equivalents: times are
69 expressed as floating point numbers, \function{time()} returns the
70 most accurate time available (using \UNIX{} \cfunction{gettimeofday()}
71 where available), and \function{sleep()} will accept a time with a
72 nonzero fraction (\UNIX{} \cfunction{select()} is used to implement
73 this, where available).
75 \item
77 The time tuple as returned by \function{gmtime()},
78 \function{localtime()}, and \function{strptime()}, and accepted by
79 \function{asctime()}, \function{mktime()} and \function{strftime()},
80 is a tuple of 9 integers:
82 \begin{tableiii}{r|l|l}{textrm}{Index}{Field}{Values}
83 \lineiii{0}{year}{(e.g.\ 1993)}
84 \lineiii{1}{month}{range [1,12]}
85 \lineiii{2}{day}{range [1,31]}
86 \lineiii{3}{hour}{range [0,23]}
87 \lineiii{4}{minute}{range [0,59]}
88 \lineiii{5}{second}{range [0,61]; see \strong{(1)} in \function{strftime()} description}
89 \lineiii{6}{weekday}{range [0,6], Monday is 0}
90 \lineiii{7}{Julian day}{range [1,366]}
91 \lineiii{8}{daylight savings flag}{0, 1 or -1; see below}
92 \end{tableiii}
94 Note that unlike the C structure, the month value is a
95 range of 1-12, not 0-11. A year value will be handled as described
96 under ``Year 2000 (Y2K) issues'' above. A \code{-1} argument as
97 daylight savings flag, passed to \function{mktime()} will usually
98 result in the correct daylight savings state to be filled in.
100 \end{itemize}
102 The module defines the following functions and data items:
105 \begin{datadesc}{accept2dyear}
106 Boolean value indicating whether two-digit year values will be
107 accepted. This is true by default, but will be set to false if the
108 environment variable \envvar{PYTHONY2K} has been set to a non-empty
109 string. It may also be modified at run time.
110 \end{datadesc}
112 \begin{datadesc}{altzone}
113 The offset of the local DST timezone, in seconds west of UTC, if one
114 is defined. This is negative if the local DST timezone is east of UTC
115 (as in Western Europe, including the UK). Only use this if
116 \code{daylight} is nonzero.
117 \end{datadesc}
119 \begin{funcdesc}{asctime}{\optional{tuple}}
120 Convert a tuple representing a time as returned by \function{gmtime()}
121 or \function{localtime()} to a 24-character string of the following form:
122 \code{'Sun Jun 20 23:21:05 1993'}. If \var{tuple} is not provided, the
123 current time as returned by \function{localtime()} is used. Note: unlike
124 the C function of the same name, there is no trailing newline.
125 \end{funcdesc}
127 \begin{funcdesc}{clock}{}
128 Return the current CPU time as a floating point number expressed in
129 seconds. The precision, and in fact the very definition of the meaning
130 of ``CPU time''\index{CPU time}, depends on that of the C function
131 of the same name, but in any case, this is the function to use for
132 benchmarking\index{benchmarking} Python or timing algorithms.
133 \end{funcdesc}
135 \begin{funcdesc}{ctime}{\optional{secs}}
136 Convert a time expressed in seconds since the epoch to a string
137 representing local time. If \var{secs} is not provided, the current time
138 as returned by \function{time()} is used. \code{ctime(\var{secs})}
139 is equivalent to \code{asctime(localtime(\var{secs}))}.
140 \end{funcdesc}
142 \begin{datadesc}{daylight}
143 Nonzero if a DST timezone is defined.
144 \end{datadesc}
146 \begin{funcdesc}{gmtime}{\optional{secs}}
147 Convert a time expressed in seconds since the epoch to a time tuple
148 in UTC in which the dst flag is always zero. If \var{secs} is not
149 provided, the current time as returned by \function{time()} is used.
150 Fractions of a second are ignored. See above for a description of the
151 tuple lay-out.
152 \end{funcdesc}
154 \begin{funcdesc}{localtime}{\optional{secs}}
155 Like \function{gmtime()} but converts to local time. The dst flag is
156 set to \code{1} when DST applies to the given time.
157 \end{funcdesc}
159 \begin{funcdesc}{mktime}{tuple}
160 This is the inverse function of \function{localtime()}. Its argument
161 is the full 9-tuple (since the dst flag is needed; use \code{-1} as
162 the dst flag if it is unknown) which expresses the time in
163 \emph{local} time, not UTC. It returns a floating point number, for
164 compatibility with \function{time()}. If the input value cannot be
165 represented as a valid time, \exception{OverflowError} is raised.
166 \end{funcdesc}
168 \begin{funcdesc}{sleep}{secs}
169 Suspend execution for the given number of seconds. The argument may
170 be a floating point number to indicate a more precise sleep time.
171 The actual suspension time may be less than that requested because any
172 caught signal will terminate the \function{sleep()} following
173 execution of that signal's catching routine. Also, the suspension
174 time may be longer than requested by an arbitrary amount because of
175 the scheduling of other activity in the system.
176 \end{funcdesc}
178 \begin{funcdesc}{strftime}{format\optional{, tuple}}
179 Convert a tuple representing a time as returned by \function{gmtime()}
180 or \function{localtime()} to a string as specified by the \var{format}
181 argument. If \var{tuple} is not provided, the current time as returned by
182 \function{localtime()} is used. \var{format} must be a string.
184 The following directives can be embedded in the \var{format} string.
185 They are shown without the optional field width and precision
186 specification, and are replaced by the indicated characters in the
187 \function{strftime()} result:
189 \begin{tableiii}{c|p{24em}|c}{code}{Directive}{Meaning}{Notes}
190 \lineiii{\%a}{Locale's abbreviated weekday name.}{}
191 \lineiii{\%A}{Locale's full weekday name.}{}
192 \lineiii{\%b}{Locale's abbreviated month name.}{}
193 \lineiii{\%B}{Locale's full month name.}{}
194 \lineiii{\%c}{Locale's appropriate date and time representation.}{}
195 \lineiii{\%d}{Day of the month as a decimal number [01,31].}{}
196 \lineiii{\%H}{Hour (24-hour clock) as a decimal number [00,23].}{}
197 \lineiii{\%I}{Hour (12-hour clock) as a decimal number [01,12].}{}
198 \lineiii{\%j}{Day of the year as a decimal number [001,366].}{}
199 \lineiii{\%m}{Month as a decimal number [01,12].}{}
200 \lineiii{\%M}{Minute as a decimal number [00,59].}{}
201 \lineiii{\%p}{Locale's equivalent of either AM or PM.}{}
202 \lineiii{\%S}{Second as a decimal number [00,61].}{(1)}
203 \lineiii{\%U}{Week number of the year (Sunday as the first day of the
204 week) as a decimal number [00,53]. All days in a new year
205 preceding the first Sunday are considered to be in week 0.}{}
206 \lineiii{\%w}{Weekday as a decimal number [0(Sunday),6].}{}
207 \lineiii{\%W}{Week number of the year (Monday as the first day of the
208 week) as a decimal number [00,53]. All days in a new year
209 preceding the first Sunday are considered to be in week 0.}{}
210 \lineiii{\%x}{Locale's appropriate date representation.}{}
211 \lineiii{\%X}{Locale's appropriate time representation.}{}
212 \lineiii{\%y}{Year without century as a decimal number [00,99].}{}
213 \lineiii{\%Y}{Year with century as a decimal number.}{}
214 \lineiii{\%Z}{Time zone name (or by no characters if no time zone exists).}{}
215 \lineiii{\%\%}{A literal \character{\%} character.}{}
216 \end{tableiii}
218 \noindent
219 Notes:
221 \begin{description}
222 \item[(1)]
223 The range really is \code{0} to \code{61}; this accounts for leap
224 seconds and the (very rare) double leap seconds.
225 \end{description}
227 Here is an example, a format for dates compatible with that specified
228 in the \rfc{822} Internet email standard.
229 \footnote{The use of \%Z is now
230 deprecated, but the \%z escape that expands to the preferred
231 hour/minute offset is not supported by all ANSI C libraries. Also,
232 a strict reading of the original 1982 \rfc{822} standard calls for
233 a two-digit year (\%y rather than \%Y), but practice moved to
234 4-digit years long before the year 2000.}
236 \begin{verbatim}
237 >>> from time import *
238 >>> strftime("%a, %d %b %Y %H:%M:%S %Z", localtime())
239 'Sat, 27 Jan 2001 05:15:05 EST'
241 \end{verbatim}
243 Additional directives may be supported on certain platforms, but
244 only the ones listed here have a meaning standardized by ANSI C.
246 On some platforms, an optional field width and precision
247 specification can immediately follow the initial \character{\%} of a
248 directive in the following order; this is also not portable.
249 The field width is normally 2 except for \code{\%j} where it is 3.
250 \end{funcdesc}
252 \begin{funcdesc}{strptime}{string\optional{, format}}
253 Parse a string representing a time according to a format. The return
254 value is a tuple as returned by \function{gmtime()} or
255 \function{localtime()}. The \var{format} parameter uses the same
256 directives as those used by \function{strftime()}; it defaults to
257 \code{"\%a \%b \%d \%H:\%M:\%S \%Y"} which matches the formatting
258 returned by \function{ctime()}. The same platform caveats apply; see
259 the local \UNIX{} documentation for restrictions or additional
260 supported directives. If \var{string} cannot be parsed according to
261 \var{format}, \exception{ValueError} is raised. Values which are not
262 provided as part of the input string are filled in with default
263 values; the specific values are platform-dependent as the XPG standard
264 does not provide sufficient information to constrain the result.
266 \strong{Note:} This function relies entirely on the underlying
267 platform's C library for the date parsing, and some of these libraries
268 are buggy. There's nothing to be done about this short of a new,
269 portable implementation of \cfunction{strptime()}.
271 Availability: Most modern \UNIX{} systems.
272 \end{funcdesc}
274 \begin{funcdesc}{time}{}
275 Return the time as a floating point number expressed in seconds since
276 the epoch, in UTC. Note that even though the time is always returned
277 as a floating point number, not all systems provide time with a better
278 precision than 1 second.
279 \end{funcdesc}
281 \begin{datadesc}{timezone}
282 The offset of the local (non-DST) timezone, in seconds west of UTC
283 (i.e. negative in most of Western Europe, positive in the US, zero in
284 the UK).
285 \end{datadesc}
287 \begin{datadesc}{tzname}
288 A tuple of two strings: the first is the name of the local non-DST
289 timezone, the second is the name of the local DST timezone. If no DST
290 timezone is defined, the second string should not be used.
291 \end{datadesc}
294 \begin{seealso}
295 \seemodule{locale}{Internationalization services. The locale
296 settings can affect the return values for some of
297 the functions in the \module{time} module.}
298 \end{seealso}