Next: , Up: System Utilities   [Contents][Index]


36.1 Timing Utilities

Octave’s core set of functions for manipulating time values are patterned after the corresponding functions from the standard C library. Several of these functions use a data structure for time that includes the following elements:

usec

Microseconds after the second (0-999999).

sec

Seconds after the minute (0-60). This number can be 60 to account for leap seconds.

min

Minutes after the hour (0-59).

hour

Hours since midnight (0-23).

mday

Day of the month (1-31).

mon

Months since January (0-11).

year

Years since 1900.

wday

Days since Sunday (0-6).

yday

Days since January 1 (0-365).

isdst

Daylight saving time flag.

gmtoff

Seconds offset from UTC.

zone

Time zone.

In the descriptions of the following functions, this structure is referred to as a tm_struct.

: seconds = time ()

Return the current time as the number of seconds since the epoch.

The epoch is referenced to 00:00:00 UTC (Coordinated Universal Time) 1 Jan 1970. For example, on Monday February 17, 1997 at 07:15:06 UTC, the value returned by time was 856163706.

See also: strftime, strptime, localtime, gmtime, mktime, now, date, clock, datenum, datestr, datevec, calendar, weekday.

: t = now ()

Return the current local date/time as a serial day number (see datenum).

The integral part, floor (now) corresponds to the number of days between today and Jan 1, 0000.

The fractional part, rem (now, 1) corresponds to the current time.

See also: clock, date, datenum.

: ctime (t)

Convert a value returned from time (or any other non-negative integer), to the local time and return a string of the same form as asctime.

The function ctime (time) is equivalent to asctime (localtime (time)). For example:

ctime (time ())
   ⇒ "Mon Feb 17 01:15:06 1997\n"

See also: asctime, time, localtime.

: tm_struct = gmtime (t)

Given a value returned from time, or any non-negative integer, return a time structure corresponding to UTC (Coordinated Universal Time).

For example:

gmtime (time ())
     ⇒ {
           usec = 0
           sec = 6
           min = 15
           hour = 7
           mday = 17
           mon = 1
           year = 97
           wday = 1
           yday = 47
           isdst = 0
           gmtoff = 0
           zone = GMT
        }

See also: strftime, strptime, localtime, mktime, time, now, date, clock, datenum, datestr, datevec, calendar, weekday.

: tm_struct = localtime (t)

Given a value returned from time, or any non-negative integer, return a time structure corresponding to the local time zone.

localtime (time ())
     ⇒ {
           usec = 0
           sec = 6
           min = 15
           hour = 1
           mday = 17
           mon = 1
           year = 97
           wday = 1
           yday = 47
           isdst = 0
           gmtoff = -21600
           zone = CST
        }

See also: strftime, strptime, gmtime, mktime, time, now, date, clock, datenum, datestr, datevec, calendar, weekday.

: seconds = mktime (tm_struct)

Convert a time structure corresponding to the local time to the number of seconds since the epoch.

For example:

mktime (localtime (time ()))
     ⇒ 856163706

See also: strftime, strptime, localtime, gmtime, time, now, date, clock, datenum, datestr, datevec, calendar, weekday.

: asctime (tm_struct)

Convert a time structure to a string using the following format: "ddd mmm mm HH:MM:SS yyyy\n".

For example:

asctime (localtime (time ()))
     ⇒ "Mon Feb 17 01:15:06 1997\n"

This is equivalent to ctime (time ()).

See also: ctime, localtime, time.

: strftime (fmt, tm_struct)

Format the time structure tm_struct in a flexible way using the format string fmt that contains ‘%’ substitutions similar to those in printf.

Except where noted, substituted fields have a fixed size; numeric fields are padded if necessary. Padding is with zeros by default; for fields that display a single number, padding can be changed or inhibited by following the ‘%’ with one of the modifiers described below. Unknown field specifiers are copied as normal characters. All other characters are copied to the output without change. For example:

strftime ("%r (%Z) %A %e %B %Y", localtime (time ()))
      ⇒ "01:15:06 AM (CST) Monday 17 February 1997"

Octave’s strftime function supports a superset of the ANSI C field specifiers.

Literal character fields:

%%

% character.

%n

Newline character.

%t

Tab character.

Numeric modifiers (a nonstandard extension):

- (dash)

Do not pad the field.

_ (underscore)

Pad the field with spaces.

Time fields:

%H

Hour (00-23).

%I

Hour (01-12).

%k

Hour (0-23).

%l

Hour (1-12).

%M

Minute (00-59).

%p

Locale’s AM or PM.

%r

Time, 12-hour (hh:mm:ss [AP]M).

%R

Time, 24-hour (hh:mm).

%s

Time in seconds since 00:00:00, Jan 1, 1970 (a nonstandard extension).

%S

Second (00-61).

%T

Time, 24-hour (hh:mm:ss).

%X

Locale’s time representation (%H:%M:%S).

%z

Offset from UTC (±hhmm), or nothing if no time zone is determinable.

%Z

Time zone (EDT), or nothing if no time zone is determinable.

Date fields:

%a

Locale’s abbreviated weekday name (Sun-Sat).

%A

Locale’s full weekday name, variable length (Sunday-Saturday).

%b

Locale’s abbreviated month name (Jan-Dec).

%B

Locale’s full month name, variable length (January-December).

%c

Locale’s date and time (Sat Nov 04 12:02:33 EST 1989).

%C

Century (00-99).

%d

Day of month (01-31).

%e

Day of month ( 1-31).

%D

Date (mm/dd/yy).

%h

Same as %b.

%j

Day of year (001-366).

%m

Month (01-12).

%U

Week number of year with Sunday as first day of week (00-53).

%w

Day of week (0-6).

%W

Week number of year with Monday as first day of week (00-53).

%x

Locale’s date representation (mm/dd/yy).

%y

Last two digits of year (00-99).

%Y

Year (1970-).

See also: strptime, localtime, gmtime, mktime, time, now, date, clock, datenum, datestr, datevec, calendar, weekday.

: [tm_struct, nchars] = strptime (str, fmt)

Convert the string str to the time structure tm_struct under the control of the format string fmt.

If fmt fails to match, nchars is 0; otherwise, it is set to the position of last matched character plus 1. Always check for this unless you’re absolutely sure the date string will be parsed correctly.

See also: strftime, localtime, gmtime, mktime, time, now, date, clock, datenum, datestr, datevec, calendar, weekday.

Most of the remaining functions described in this section are not patterned after the standard C library. Some are available for compatibility with MATLAB and others are provided because they are useful.

: clock ()

Return the current local date and time as a date vector.

The date vector contains the following fields: current year, month (1-12), day (1-31), hour (0-23), minute (0-59), and second (0-61). The seconds field has a fractional part after the decimal point for extended accuracy.

For example:

fix (clock ())
  ⇒   1993      8     20      4     56      1

clock is more accurate on systems that have the gettimeofday function.

See also: now, date, datevec.

: date ()

Return the current date as a character string in the form DD-MMM-YYYY.

For example:

date ()
  ⇒ 20-Aug-1993

See also: now, clock, datestr, localtime.

: etime (t2, t1)

Return the difference in seconds between two time values returned from clock (t2 - t1).

For example:

t0 = clock ();
# many computations later…
elapsed_time = etime (clock (), t0);

will set the variable elapsed_time to the number of seconds since the variable t0 was set.

See also: tic, toc, clock, cputime, addtodate.

: [total, user, system] = cputime ();

Return the CPU time used by your Octave session.

The first output is the total time spent executing your process and is equal to the sum of second and third outputs, which are the number of CPU seconds spent executing in user mode and the number of CPU seconds spent executing in system mode, respectively.

If your system does not have a way to report CPU time usage, cputime returns 0 for each of its output values.

Note that because Octave used some CPU time to start, it is reasonable to check to see if cputime works by checking to see if the total CPU time used is nonzero.

See also: tic, toc.

: is_leap_year ()
: is_leap_year (year)

Return true if year is a leap year and false otherwise.

If no year is specified, is_leap_year uses the current year.

For example:

is_leap_year (2000)
   ⇒ 1

See also: weekday, eomday, calendar.

: tic ()
: id = tic ()

Initialize a wall-clock timer.

Calling tic without an output argument resets the internal timer. Subsequent calls to toc return the number of seconds since the timer was set.

If called with one output argument, tic creates a new timer instance and returns a timer identifier id. The id is a scalar of type uint64 that may be passed to toc to check elapsed time on this timer, rather than the default internal timer.

Example 1 : benchmarking code with internal timer

tic;
# many computations later…
elapsed_time = toc;

Example 2 : mixed timer id and internal timer

tic;
pause (1);
toc
⇒ Elapsed time is 1.0089 seconds.
id = tic;
pause (2);
toc (id)
⇒ Elapsed time is 2.01142 seconds.
toc
Elapsed time is 3.02308 seconds.

Calling tic and toc in this way allows nested timing calls.

If you are more interested in the CPU time that your process used, you should use the cputime function instead. The tic and toc functions report the actual wall clock time that elapsed between the calls. This may include time spent processing other jobs or doing nothing at all.

See also: toc, cputime.

: toc ()
: toc (id)
: elapsed_time = toc (…)

Measure elapsed time on a wall-clock timer.

With no arguments, return the number of seconds elapsed on the internal timer since the last call to tic.

When given the identifier id of a specific timer, return the number of seconds elapsed since the timer id was initialized.

See tic, for examples of the use of tic/toc.

See also: tic, cputime.

: pause ()
: pause (n)
: old_state = pause ("on")
: old_state = pause ("off")
: old_state = pause ("query")

Suspend the execution of the program or change the state of the pause function.

If invoked without an input arguments then the program is suspended until a character is typed. If argument n is a positive real value, it indicates the number of seconds the program shall be suspended, for example:

tic; pause (0.05); toc
     -| Elapsed time is 0.05039 seconds.

The following example prints a message and then waits 5 seconds before clearing the screen.

disp ("wait please...");
pause (5);
clc;

If invoked with a string argument "on", "off", or "query", the state of the pause function is changed or queried. When the state is "off", the pause function returns immediately. The optional return value contains the previous state of the pause function. In the following example pause is disabled locally:

old_state = pause ("off");
tic; pause (0.05); toc
     -| Elapsed time is 3.00407e-05 seconds.
pause (old_state);

While the program is suspended Octave still handles figures painting and graphics callbacks execution.

See also: kbhit.

: days = datenum (datevec)
: days = datenum (year, month, day)
: days = datenum (year, month, day, hour)
: days = datenum (year, month, day, hour, minute)
: days = datenum (year, month, day, hour, minute, second)
: days = datenum ("datestr")
: days = datenum ("datestr", f)
: days = datenum ("datestr", p)
: [days, secs] = datenum (…)

Return the date/time input as a serial day number, with Jan 1, 0000 defined as day 1.

The integer part, floor (days) counts the number of complete days in the date input.

The fractional part, rem (days, 1) corresponds to the time on the given day.

The input may be a date vector (see datevec), datestr (see datestr), or directly specified as input.

When processing input datestrings, f is the format string used to interpret date strings (see datestr). If no format f is specified, then a relatively slow search is performed through various formats. It is always preferable to specify the format string f if it is known. Formats which do not specify a particular time component will have the value set to zero. Formats which do not specify a date will default to January 1st of the current year.

p is the year at the start of the century to which two-digit years will be referenced. If not specified, it defaults to the current year minus 50.

The optional output secs holds the time on the specified day with greater precision than days.

Notes:

Caution: datenums represent a specific time for the Earth as a a whole. They do not take in to account time zones (shifts in time based on location), nor seasonal changes due to Daylight Savings Time (shifts in time based on local regulation). Be aware that it is possible to create datenums that, when interpreted by a function which accounts for time zone and DST shifts such as datestr, are nonexistent or ambiguous.

Caution: this function does not attempt to handle Julian calendars so dates before October 15, 1582 are wrong by as much as eleven days. Also, be aware that only Roman Catholic countries adopted the calendar in 1582. It took until 1924 for it to be adopted everywhere. See the Wikipedia entry on the Gregorian calendar for more details.

Warning: leap seconds are ignored. A table of leap seconds is available on the Wikipedia entry for leap seconds.

See also: datestr, datevec, now, clock, date.

: str = datestr (date)
: str = datestr (date, f)
: str = datestr (date, f, p)

Format the given date/time according to the format f and return the result in str.

date is a serial date number (see datenum), a date vector (see datevec), or a string or cell array of strings. In the latter case, it is passed to datevec to guess the input date format.

f can be an integer which corresponds to one of the codes in the table below, or a date format string.

p is the year at the start of the century in which two-digit years are to be interpreted in. If not specified, it defaults to the current year minus 50.

For example, the date 730736.65149 (2000-09-07 15:38:09.0934) would be formatted as follows:

CodeFormatExample
0dd-mmm-yyyy HH:MM:SS07-Sep-2000 15:38:09
1dd-mmm-yyyy07-Sep-2000
2mm/dd/yy09/07/00
3mmmSep
4mS
5mm09
6mm/dd09/07
7dd07
8dddThu
9dT
10yyyy2000
11yy00
12mmmyySep00
13HH:MM:SS15:38:09
14HH:MM:SS PM3:38:09 PM
15HH:MM15:38
16HH:MM PM3:38 PM
17QQ-YYQ3-00
18QQQ3
19dd/mm07/09
20dd/mm/yy07/09/00
21mmm.dd,yyyy HH:MM:SSSep.07,2000 15:38:08
22mmm.dd,yyyySep.07,2000
23mm/dd/yyyy09/07/2000
24dd/mm/yyyy07/09/2000
25yy/mm/dd00/09/07
26yyyy/mm/dd2000/09/07
27QQ-YYYYQ3-2000
28mmmyyyySep2000
29yyyy-mm-dd2000-09-07
30yyyymmddTHHMMSS20000907T153808
31yyyy-mm-dd HH:MM:SS2000-09-07 15:38:08

If f is a format string, the following symbols are recognized:

SymbolMeaningExample
yyyyFull year2005
yyTwo-digit year05
mmmmFull month nameDecember
mmmAbbreviated month nameDec
mmNumeric month number (padded with zeros)01, 08, 12
mFirst letter of month name (capitalized)D
ddddFull weekday nameSunday
dddAbbreviated weekday nameSun
ddNumeric day of month (padded with zeros)11
dFirst letter of weekday name (capitalized)S
HHHour of day, padded with zeros,09:00
or padded with spaces if PM is set9:00 AM
MMMinute of hour (padded with zeros)10:05
SSSecond of minute (padded with zeros)10:05:03
FFFMilliseconds of second (padded with zeros)10:05:03.012
AMUse 12-hour time format11:30 AM
PMUse 12-hour time format11:30 PM

If f is not specified or is -1, then use 0, 1 or 16, depending on whether the date portion or the time portion of date is empty.

If p is nor specified, it defaults to the current year minus 50.

If a matrix or cell array of dates is given, a column vector of date strings is returned.

See also: datenum, datevec, date, now, clock.

: v = datevec (date)
: v = datevec (date, f)
: v = datevec (date, p)
: v = datevec (date, f, p)
: [y, m, d, h, mi, s] = datevec (…)

Convert a serial date number (see datenum) or date string (see datestr) into a date vector.

A date vector is a row vector with six members, representing the year, month, day, hour, minute, and seconds respectively.

f is the format string used to interpret date strings (see datestr). If date is a string, but no format is specified, then a relatively slow search is performed through various formats. It is always preferable to specify the format string f if it is known. Formats which do not specify a particular time component will have the value set to zero. Formats which do not specify a date will default to January 1st of the current year.

p is the year at the start of the century to which two-digit years will be referenced. If not specified, it defaults to the current year minus 50.

See also: datenum, datestr, clock, now, date.

: d = addtodate (d, q, f)

Add q amount of time (with units f) to the serial datenum, d.

f must be one of "year", "month", "day", "hour", "minute", "second", or "millisecond".

See also: datenum, datevec, etime.

: c = calendar ()
: c = calendar (d)
: c = calendar (y, m)
: calendar (…)

Return the current monthly calendar in a 6x7 matrix.

If d is specified, return the calendar for the month containing the date d, which must be a serial date number or a date string.

If y and m are specified, return the calendar for year y and month m.

If no output arguments are specified, print the calendar on the screen instead of returning a matrix.

See also: datenum, datestr.

: [n, s] = weekday (d)
: [n, s] = weekday (d, format)

Return the day of the week as a number in n and as a string in s.

The days of the week are numbered 1–7 with the first day being Sunday.

d is a serial date number or a date string.

If the string format is not present or is equal to "short" then s will contain the abbreviated name of the weekday. If format is "long" then s will contain the full name.

Table of return values based on format:

n"short""long"
1SunSunday
2MonMonday
3TueTuesday
4WedWednesday
5ThuThursday
6FriFriday
7SatSaturday

See also: eomday, is_leap_year, calendar, datenum, datevec.

: e = eomday (y, m)

Return the last day of the month m for the year y.

See also: weekday, datenum, datevec, is_leap_year, calendar.

: datetick ()
: datetick (axis_str)
: datetick (date_format)
: datetick (axis_str, date_format)
: datetick (…, "keeplimits")
: datetick (…, "keepticks")
: datetick (hax, …)

Add date-formatted tick labels to an axis.

The axis to apply the ticks to is determined by axis_str which can take the values "x", "y", or "z". The default value is "x".

The formatting of the labels is determined by the variable date_format, which can either be a string or positive integer that datestr accepts.

If the first argument hax is an axes handle, then plot into this axes, rather than the current axes returned by gca.

See also: datenum, datestr.


Next: , Up: System Utilities   [Contents][Index]