datetime-fortran is a time and date manipulation library for Fortran. It provides classes for date and time (datetime), and time difference representation (timedelta) as well as arithmetic and comparison operators and associated methods for their manipulation. It also provides an interface to C/C++ tm struct, and associated strftime and strptime functions. Since version 0.2.0, also provides a clock class. datetime-fortran came about due to the lack of time handling facilities in standard Fortran language. It is freely available under the GNU General Public License version 3. Contact me if you want to use datetime-fortran in a non-GPL software. Please send suggestions and bug reports by e-mail or through this Github page. See the list of current issues if you would like to contribute to the code.
-
Arithmetic operators
+and-for datetime and timedelta objects; -
Comparison operators
>,>=,<,<=,==and/=for datetime and timedelta objects; -
Basic timezone handling and arithmetic;
-
Interfaces to C/C++ routines c_strftime and c_strptime through
ISO_C_BINDING. Since version0.3.0, datetime-bound method strftime and function strptime that return a datetime instance are available. -
Lightweight and portable;
-
Free to modify and distribute under the terms of the GNU General Public License version 3.
-
Tested with the following compilers:
- gfortran 4.7.2, 4.8.2
- ifort 13.1.1, 14.0.2, 15.0.0
- pgf90 13.6-0, 13.10-0 (thanks to Timothy Hilton)
- xlf 14.1.0.5 (thanks to Bjoern Hendrik Fock)
datetime-fortran library provides the following derived types:
datetime, timedelta,
clock and tm_struct.
Main date and time object, defined as:
TYPE :: datetime
! COMPONENTS:
INTEGER :: year = 1 ! Year [1-HUGE(year)]
INTEGER :: month = 1 ! Month in year [1-12]
INTEGER :: day = 1 ! Day in month [1-31]
INTEGER :: hour = 0 ! Hour in day [0-23]
INTEGER :: minute = 0 ! Minute in hour [0-59]
INTEGER :: second = 0 ! Second in minute [0-59]
INTEGER :: millisecond = 0 ! Milliseconds in second [0-999]
REAL :: tz = 0 ! Timezone offset from UTC [hours]
CONTAINS
! METHODS:
PROCEDURE :: addMilliseconds
PROCEDURE :: addSeconds
PROCEDURE :: addMinutes
PROCEDURE :: addHours
PROCEDURE :: addDays
PROCEDURE :: isocalendar
PROCEDURE :: isoformat
PROCEDURE :: isValid
PROCEDURE :: now
PROCEDURE :: secondsSinceEpoch
PROCEDURE :: strftime
PROCEDURE :: tm
PROCEDURE :: tzOffset
PROCEDURE :: utc
PROCEDURE :: weekday
PROCEDURE :: weekdayLong
PROCEDURE :: weekdayShort
PROCEDURE :: yearday
ENDTYPE datetimedatetime components are initialized by default, so all arguments are optional. Arguments may be provided as positional arguments, in the order of their declaration, or as keyword arguments, in any order. If both positional and keyword arguments are used, no positional arguments may appear after a keyword argument.
USE datetime_module,ONLY:datetime
TYPE(datetime) :: a
! Initialize as default:
a = datetime() ! 0001-01-01 00:00:00
! Components can be specified by position:
a = datetime(1984,12,10) ! 1984-12-10 00:00:00
! Or by keyword:
a = datetime(month=1,day=1,year=1970) ! 1970-01-01 00:00:00
! Or combined:
a = datetime(2013,2,minute=23,day=5) ! 2013-02-05 00:23:00
! With timezone offset:
a = datetime(2013,2,minute=23,day=5,tz=-4) ! 2013-02-05 00:23:00 -0400
! Do not use positional after keyword arguments:
a = datetime(year=2013,2,minute=23,day=5) ! ILLEGALNote that the current implementation of datetime does not support daylight saving time (DST) information.
PURE ELEMENTAL SUBROUTINE addMilliseconds(self,ms)
! ARGUMENTS:
CLASS(datetime),INTENT(INOUT) :: self
INTEGER, INTENT(IN) :: ms ! Number of milliseconds to addUsed internally by binary arithmetic operators + and - when adding/subtracting a timedelta instance to/from a datetime instance. In general, there is no need to use this method from external programs. However, it may be convenient and create less overhead if the operation needs to be performed on a large array of datetime instances.
ms Integer number of milliseconds to add. May be negative for subtraction.
None
USE datetime_module,ONLY:datetime
TYPE(datetime) :: a
! Initialize:
a = datetime(2013,1,1,0,0,0,0) ! 2013-01-01 00:00:00.000
! Add:
CALL a%addMilliseconds(100) ! a becomes: 2013-01-01 00:00:00.100PURE ELEMENTAL SUBROUTINE addSeconds(self,s)
! ARGUMENTS:
CLASS(datetime),INTENT(INOUT) :: self
INTEGER, INTENT(IN) :: s ! Number of seconds to addUsed internally by binary arithmetic operators + and - when adding/subtracting a timedelta instance to/from a datetime instance. In general, there is no need to use this method from external programs. However, it may be convenient and create less overhead if the operation needs to be performed on a large array of datetime instances.
s Integer number of seconds to add. May be negative for subtraction.
None
USE datetime_module,ONLY:datetime
TYPE(datetime) :: a
! Initialize:
a = datetime(2013,1,1,0,0,0,0) ! 2013-01-01 00:00:00.000
! Add:
CALL a%addSeconds(10) ! a becomes: 2013-01-01 00:00:10.000PURE ELEMENTAL SUBROUTINE addMinutes(self,m)
! ARGUMENTS:
CLASS(datetime),INTENT(INOUT) :: self
INTEGER, INTENT(IN) :: m ! Number of minutes to addUsed internally by binary arithmetic operators + and - when adding/subtracting a timedelta instance to/from a datetime instance. In general, there is no need to use this method from external programs. However, it may be convenient and create less overhead if the operation needs to be performed on a large array of datetime instances.
m Integer number of minutes to add. May be negative for subtraction.
None
USE datetime_module,ONLY:datetime
TYPE(datetime) :: a
! Initialize:
a = datetime(2013,1,1,0,0,0,0) ! 2013-01-01 00:00:00.000
! Add:
CALL a%addMinutes(10) ! a becomes: 2013-01-01 00:10:00.000PURE ELEMENTAL SUBROUTINE addHours(self,h)
! ARGUMENTS:
CLASS(datetime),INTENT(INOUT) :: self
INTEGER, INTENT(IN) :: h ! Number of hours to addUsed internally by binary arithmetic operators + and - when adding/subtracting a timedelta instance to/from a datetime instance. In general, there is no need to use this method from external programs. However, it may be convenient and create less overhead if the operation needs to be performed on a large array of datetime instances.
h Integer number of hours to add. May be negative for subtraction.
None
USE datetime_module,ONLY:datetime
TYPE(datetime) :: a
! Initialize:
a = datetime(2013,1,1,0,0,0,0) ! 2013-01-01 00:00:00.000
! Add:
CALL a%addHours(12) ! a becomes: 2013-01-01 12:00:00.000PURE ELEMENTAL SUBROUTINE addDays(self,d)
! ARGUMENTS:
CLASS(datetime),INTENT(INOUT) :: self
INTEGER, INTENT(IN) :: d ! Number of days to addUsed internally by binary arithmetic operators + and - when adding/subtracting a timedelta instance to/from a datetime instance. In general, there is no need to use this method from external programs. However, it may be convenient and create less overhead if the operation needs to be performed on a large array of datetime instances.
d Integer number of days to add. May be negative for subtraction.
None
USE datetime_module,ONLY:datetime
TYPE(datetime) :: a
! Initialize:
a = datetime(2013,1,1,0,0,0,0) ! 2013-01-01 00:00:00.000
! Add:
CALL a%addDays(7) ! a becomes: 2013-01-08 00:00:00.000FUNCTION isocalendar(self)
! ARGUMENTS:
CLASS(datetime),INTENT(IN) :: self
INTEGER,DIMENSION(3) :: isocalendarReturns an array of 3 integers: year, week number, and week day, as defined by ISO 8601 week date. The ISO calendar is a widely used variant of the Gregorian calendar. The ISO year consists of 52 or 53 full weeks. A week starts on a Monday (1) and ends on a Sunday (7). The first week of an ISO year is the first (Gregorian) calendar week of a year containing a Thursday. This is called week number 1, and the ISO year of that Thursday is the same as its Gregorian year.
datetime%isocalendar() is equivalent to Python's datetime.datetime.isocalendar().
None
isocalendar A rank 1 integer array of length 3. Contains year, week number
and week day.
USE datetime_module,ONLY:datetime
TYPE(datetime) :: a
a = datetime(2013,1,1)
WRITE(*,*)a%isocalendar() ! Prints: 2013 1 2PURE ELEMENTAL CHARACTER(LEN=23) FUNCTION isoformat(self,sep)
! ARGUMENTS:
CLASS(datetime), INTENT(IN) :: self
CHARACTER(LEN=1),INTENT(IN),OPTIONAL :: sepReturns a character string of length 23 that contains date and time in ISO 8601 format.
datetime%isoformat() is equivalent to Python's datetime.datetime.isoformat(), with the only difference being that datetime%isoformat() returns the milliseconds at the end of the string, where as datetime.datetime.isoformat() returns microseconds.
sep is an optional argument that specifies which character of length 1 will
separate date and time entries. If ommited, defaults to T.
USE datetime_module,ONLY:datetime
TYPE(datetime) :: a
a = datetime(1984,12,10,13,5,0)
! Without arguments:
WRITE(*,*)a%isoformat() ! Prints 1984-12-10T13:05:00.000
! With a specified separator:
WRITE(*,*)a%isoformat(' ') ! Prints 1984-12-10 13:05:00.000PURE ELEMENTAL LOGICAL FUNCTION isValid(self)
! ARGUMENTS:
CLASS(datetime),INTENT(IN) :: selfReturns .TRUE. if all datetime instance components
have valid values, and .FALSE. otherwise. Components have valid values
if they are within the range indicated in datetime
derived type description.
Useful for debugging and validating user input.
USE datetime_module,ONLY:datetime
TYPE(datetime) :: a
a = datetime(1984,12,10,13,5,0)
WRITE(*,*)a%isValid() ! .TRUE.
a = datetime(1984,13,10,13,5,0)
WRITE(*,*)a%isValid() ! .FALSE.TYPE(datetime) FUNCTION now(self)
! ARGUMENTS:
CLASS(datetime),INTENT(IN) :: selfReturns the datetime instance representing the current machine time. Does not support timezones.
None.
self A datetime instance with current machine time.
USE datetime_module,ONLY:datetime
TYPE(datetime) :: a
a = a%now() ! Assigns current machine time to aINTEGER FUNCTION secondsSinceEpoch(self)
! ARGUMENTS
CLASS(datetime),INTENT(IN) :: selfReturns an integer number of seconds since the
UNIX Epoch, 1970-01-01 00:00:00 +0000 (UTC).
None.
secondsSinceEpoch An INTEGER scalar containing number of seconds since
UNIX Epoch.
USE datetime_module,ONLY:datetime
TYPE(datetime) :: a
! Initialize:
a = datetime(2013,1,1)
WRITE(*,*)a%secondsSinceEpoch() CHARACTER(LEN=MAXSTRLEN) FUNCTION strftime(self,format)
! ARGUMENTS:
CLASS(datetime), INTENT(IN) :: self
CHARACTER(LEN=*),INTENT(IN) :: formatA datetime-bound method that serves as a wrapper around the
C routine strftime.
datetime%strftime takes only the format string as argument,
and returns the character string representation of the time
information contained in the datetime instance. Thus, this function
takes care of the conversion to tm_struct and calling the raw C strftime.
Because Fortran does not allow assumed-length character strings as
the type of the function result, a fixed length of MAXSTRLEN is used.
MAXSTRLEN is currently set to 99. It is assumed that the desired
time string is shorter than this value.
Any resulting string shorter than MAXSTRLEN is padded with spaces,
so it is best to trim the result using the TRIM intrinsic function
(see the usage example below).
This datetime-bound method is available since version 0.3.0.
format A character string describing the desired format of date and time.
Same as the format for the raw C strftime.
A CHARACTER(LEN=MAXSTRLEN) representation of datetime using format.
USE datetime_module
TYPE(datetime) :: a
a = a % now()
WRITE(*,*)a%isoformat()
WRITE(*,*)TRIM(a%strftime("%Y %B %d"))PURE ELEMENTAL TYPE(tm_struct) FUNCTION tm(self)
! ARGUMENTS:
CLASS(datetime),INTENT(IN) :: selfReturns a tm_struct instance that matches the time and date information in the caller datetime instance.
USE datetime_module,ONLY:datetime
TYPE(datetime) :: a
TYPE(tm_struct) :: tm
! Initialize:
a = datetime(2013,1,1)
! Get tm_struct from datetime:
tm = a%tm()PURE ELEMENTAL CHARACTER(LEN=5) FUNCTION tzOffset(self)
! ARGUMENTS:
CLASS(datetime),INTENT(IN) :: selfGiven a datetime instance, returns a character string with timezone
offset in hours from UTC (Coordinated Universal Time), in format +hhmm
or -hhmm, depending on the sign, where hh are hours and mm are minutes.
None.
tzOffset A CHARACTER(LEN=5) in the form +hhmm
or -hhmm, depending on the sign.
USE datetime_module,ONLY:datetime
TYPE(datetime) :: a
TYPE(tm_struct) :: tm
! Initialize a datetime instance with timezone offset of -4.75 hours:
a = datetime(2013,1,1,tz=-4.75)
! Write tzOffset on screen:
WRITE(*,*)a%tzOffset ! -0445 (offset of 4 hours and 45 minutes)PURE ELEMENTAL TYPE(datetime) FUNCTION utc(self)
! ARGUMENTS:
CLASS(datetime),INTENT(IN) :: selfReturns the datetime instance at Coordinated Universal Time (UTC).
None.
utc A datetime instance with at UTC (tz = 0).
USE datetime_module,ONLY:datetime
TYPE(datetime) :: a
TYPE(tm_struct) :: tm
! Initialize a datetime instance with timezone offset of -4.75 hours:
a = datetime(2013,1,1,tz=-4.75)
WRITE(*,*)a%isoformat()//a%tzOffset() ! 2013-01-01T00:00:00.000-0445
! Convert a to UTC:
a = a%utc()
WRITE(*,*)a%isoformat()//a%tzOffset() ! 2013-01-01T04:45:00.000+0000PURE ELEMENTAL INTEGER FUNCTION weekday(self)
CLASS(datetime),INTENT(IN) :: selfA datetime-bound method to calculate day of the week using Zeller's congruence. Returns an integer scalar in the range of [0-6], starting from Sunday.
USE datetime_module,ONLY:datetime
TYPE(datetime) :: a
! Initialize:
a = datetime(2013,1,1)
WRITE(*,*)a%weekday() ! 2PURE ELEMENTAL CHARACTER(LEN=9) FUNCTION weekdayLong(self)
CLASS(datetime),INTENT(IN) :: self
Returns the full name of the day of the week.
USE datetime_module,ONLY:datetime
TYPE(datetime) :: a
! Initialize:
a = datetime(2013,1,1)
WRITE(*,*)a%weekdayLong() ! TuesdayPURE ELEMENTAL CHARACTER(LEN=3) FUNCTION weekdayShort(self)
CLASS(datetime),INTENT(IN) :: self
Returns the abbreviated (e.g. Mon) name of the day of the week.
USE datetime_module,ONLY:datetime
TYPE(datetime) :: a
! Initialize:
a = datetime(2013,1,1)
WRITE(*,*)a%weekdayShort() ! TuePURE ELEMENTAL INTEGER FUNCTION yearday(self)
CLASS(datetime),INTENT(IN) :: selfdatetime-bound procedure.
Returns integer day of the year (ordinal date).
Equals to 1 for any January 1, 365 for a December 31 on a non-leap year,
and 366 for a December 31 on a leap year.
USE datetime_module,ONLY:datetime
TYPE(datetime) :: a
! Initialize:
a = datetime(2013,5,1)
WRITE(*,*)a%yearday() ! 121 Represents a duration of time, and a difference between two datetime objects. It is defined as:
TYPE :: timedelta
! COMPONENTS:
INTEGER :: days = 0
INTEGER :: hours = 0
INTEGER :: minutes = 0
INTEGER :: seconds = 0
INTEGER :: milliseconds = 0
CONTAINS
! METHODS:
PROCEDURE :: total_seconds
ENDTYPE timedeltaAll arguments are optional and default to 0. Similarly to datetime objects, timedelta instances can be initialized using positional and/or keyword arguments. In addition, a timedelta object is a result of subtraction between two datetime objects.
USE datetime_module
TYPE(datetime) :: a,b
TYPE(timedelta) :: c
! Initialize as default
c = timedelta()
! Positional arguments:
c = timedelta(0,1,15,0,0) ! 1 hour and 15 minutes
! Keyword arguments:
c = timedelta(days=1,hours=12) ! 1 day and 12 hours
! Difference between datetimes:
a = datetime(2013,5,12,22,0,0) ! 2013-05-12 22:00:00
b = datetime(2012,9,18,14,0,0) ! 2012-09-18 14:00:00
! Subtract to get timedelta:
c = a-b PURE ELEMENTAL REAL(KIND=real_dp) FUNCTION total_seconds(self)
! ARGUMENTS:
CLASS(timedelta),INTENT(IN) :: selfA timedelta-bound method that returns a number of seconds contained in the time interval defined by the timedelta instance. This method is equivalent to Python's datetime.timedelta.total_seconds function.
None
total_seconds A total number of seconds (of type REAL(KIND=real_dp))
contained in the timedelta instance.
USE datetime_module,ONLY:timedelta
TYPE(timedelta) :: td
td = timedelta(days=5,hours=12,minutes=15,seconds=7,milliseconds=123)
WRITE(*,*)td%total_seconds() ! 476107.12300000002A generic clock object that contains start and stop times, tick increment and reset and tick methods. Most useful when needing to keep track of many datetime instances that change at different rates, for example, physical models with different time steps.
Definition:
TYPE :: clock
! COMPONENTS:
TYPE(datetime) :: startTime = datetime()
TYPE(datetime) :: stopTime = datetime()
TYPE(datetime) :: currentTime = datetime()
TYPE(timedelta) :: tickInterval = timedelta()
LOGICAL :: alarm = .FALSE.
! Clock status flags
LOGICAL :: started = .FALSE.
LOGICAL :: stopped = .FALSE.
CONTAINS
! METHODS:
PROCEDURE :: reset
PROCEDURE :: tick
ENDTYPE clockclock components are initialized by default, and all arguments
are optional. However, a clock instance must be initialized
with some sane values of clock%startTime, clock%stopTime and clock%tickIncrement
in order to be useful.
USE datetime_module
TYPE(clock) :: myClock
TYPE(datetime) :: myTime
! Initialize myTime
myTime = myTime%now()
! Initialize myClock
! Starts from myTime, stops 1 hour later, 1 minute per tick
myClock = clock(startTime = myTime, &
stopTime = myTime+timedelta(hours=1),&
tickInterval = timedelta(minutes=1))
DO
CALL myClock % tick()
! Report current time after each tick
WRITE(*,*)myClock % currentTime % isoformat(' ')
! If clock has reached stopTime, exit loop
IF(myClock % stopped)THEN
EXIT
ENDIF
ENDDOPURE ELEMENTAL SUBROUTINE reset(self)
! ARGUMENTS:
CLASS(clock),INTENT(INOUT) :: selfResets the clock to its start time.
None
None
CALL myClock%reset() ! Resets myClock%currentTime to myClock%startTimePURE ELEMENTAL SUBROUTINE tick(self)
! ARGUMENTS:
CLASS(clock),INTENT(INOUT) :: selfIncrements the currentTime of the clock instance by one tickInterval.
Sets the clock%stopped flag to .TRUE. if clock%currentTime equals
or exceeds clock%stopTime.
None
None
See clock for an example.
Time object compatible with C/C++ tm struct. Available mainly for the purpose of calling c_strftime and c_strptime procedures.
TYPE,BIND(c) :: tm_struct
! COMPONENTS:
INTEGER(KIND=c_int) :: tm_sec ! Seconds [0-60] (1 leap second)
INTEGER(KIND=c_int) :: tm_min ! Minutes [0-59]
INTEGER(KIND=c_int) :: tm_hour ! Hours [0-23]
INTEGER(KIND=c_int) :: tm_mday ! Day [1-31]
INTEGER(KIND=c_int) :: tm_mon ! Month [0-11]
INTEGER(KIND=c_int) :: tm_year ! Year - 1900
INTEGER(KIND=c_int) :: tm_wday ! Day of week [0-6]
INTEGER(KIND=c_int) :: tm_yday ! Days in year [0-365]
INTEGER(KIND=c_int) :: tm_isdst ! DST [-1/0/1]
! METHODS: None.
ENDTYPE tm_structThe datetime-fortran library provides arithmetic and comparison operators for datetime and timedelta objects.
Addition (+) and subtraction (-) operators are available for
the following combination of derived type pairs:
-
datetime + timedelta, returns adatetimeinstance; -
timedelta + datetime, returns adatetimeinstance; -
timedelta + timedelta, returns atimedeltainstance; -
timedelta - timedelta, returns atimedeltainstance; -
datetime - datetime, returns atimedeltainstance; -
-timedelta(unary minus), returns atimedeltainstance.
Note that datetime - datetime operation accounts for timezone (tz)
offsets in each of the datetime instances.
The resulting timedeltathus includes the difference between timezones.
datetime-fortran supports following binary comparison operators for
datetime and timedelta objects:
==, /=, >, >=, < and <=.
FUNCTION c_strftime(str,slen,format,tm)BIND(c,name='strftime')RESULT(rc)
! ARGUMENTS:
CHARACTER(KIND=c_char),DIMENSION(*),INTENT(OUT) :: str
INTEGER(KIND=c_int),VALUE, INTENT(IN) :: slen
CHARACTER(KIND=c_char),DIMENSION(*),INTENT(IN) :: format
TYPE(tm_struct), INTENT(IN) :: tm
INTEGER(KIND=c_int) :: rcAn interface to a C/C++ standard library routine.
Copies into str the content of format, expanding its format specifiers
into the corresponding values that represent the time described in tm,
with a limit of slen characters.
Note: This function was renamed from strftime to c_strftime in version 0.3.0 to avoid name conflict with datetime-bound method strftime. If working with datetime instances, use datetime%strftime instead.
str is the destination character string with the requested date and time.
slen is the maximum number of characters to be copied to str,
including the terminating null-character, CHAR(0).
format is the character string containing any combination of regular characters and special format specifiers.
These format specifiers are replaced by the function to the corresponding values to represent the time specified in tm.
For more information on format specifiers see http://www.cplusplus.com/reference/ctime/strftime/.
tm is an instance of the type tm_struct, containing date and time values to be processed.
If the resulting string fits in less than slen characters including the terminating null-character,
the total number of characters copied to str (not including the terminating null-character) is returned.
Otherwise, zero is returned and the contents of the array are indeterminate.
USE datetime_module
TYPE(datetime) :: a
CHARACTER(LEN=20) :: res
INTEGER :: rc
a = a % now()
rc = c_strftime(res,20,"%Y %B %d"//CHAR(0),a%tm())FUNCTION c_strptime(str,format,tm)BIND(c,name='strptime')RESULT(rc)
CHARACTER(KIND=c_char),DIMENSION(*),INTENT(IN) :: str
CHARACTER(KIND=c_char),DIMENSION(*),INTENT(IN) :: format
TYPE(tm_struct), INTENT(OUT) :: tm
CHARACTER(KIND=c_char,LEN=1) :: rcAn interface to a C/C++ standard library routine.
Converts the character string str to values which are stored in tm, using the format specified by format.
Note: This function was renamed from strptime to c_strptime in version 0.3.0 to avoid name conflicts with strptime which operates on datetime instances. If working with datetime instances, use strptime instead.
str is the character string containing date and time information.
format is the character string containing any combination of regular characters and special format specifiers,
describing the date and time information in str.
tm is an instance of the type tm_struct, in which the date and time values will be filled upon successful completion
of the c_strptime function.
Upon successful completion, c_strptime returns the character following the last character parsed. Otherwise, a null character is returned.
Extracting time difference between two time strings using c_strptime and tm2date:
USE datetime_module
TYPE(datetime) :: date1,date2
TYPE(tm_struct) :: ctime
TYPE(timedelta) :: timediff
! Return code for strptime
CHARACTER(LEN=1) :: rc
! Example times in "YYYYMMDD hhmmss" format
CHARACTER(LEN=15) :: str1 = "20130512 091519"
CHARACTER(LEN=15) :: str2 = "20131116 120418"
! Get tm_struct instance from str1
rc = c_strptime(str1,"%Y%m%d %H%M%S"//CHAR(0),ctime)
date1 = tm2date(ctime)
! Get tm_struct instance from str2
rc = c_strptime(str2,"%Y%m%d %H%M%S"//CHAR(0),ctime)
date2 = tm2date(ctime)
timediff = date2-date1
WRITE(*,*)timediff
WRITE(*,*)timediff%total_seconds()This example outputs the following:
188 2 48 58 1000
16253339.0000000
PURE ELEMENTAL REAL(KIND=real_dp) FUNCTION date2num(d)
! ARGUMENTS:
TYPE(datetime),INTENT(IN) :: dReturns the number of days since 0001-01-01 00:00:00 UTC,
given a datetime instance d.
This function is similar in what it returns to analogous functions
in Python (matplotlib.dates.date2num)
and MATLAB's datenum.
Note that matplotlib.dates.date2num returns the number of days since 0001-01-01 00:00:00 UTC plus 1
(for historical reasons),
and MATLAB's datenum
returns the number of days since 0000-01-01 00:00:00 UTC.
In datetime-fortran, we choose the reference time of 0001-01-01 00:00:00 UTC
as we consider it to be the least astonishing for the average user.
Thus, MATLAB and Python users should be cautious when using
datetime-fortran's date2num() function.
date2num is the inverse function of num2date,
so by definition, a == num2date(date2num(a)) evaluates as .TRUE.
for any datetime instance a.
Similarly, b == date2num(num2date(b)) evaluates as .TRUE.
for any variable b of type REAL(KIND=real_dp).
d A datetime instance.
date2num A REAL(KIND=real_dp) number of days since 0001-01-01 00:00:00 UTC.
real_dp is defined as:
INTEGER,PARAMETER :: real_dp = KIND(1d0)USE datetime_module,ONLY:datetime,date2num
TYPE(datetime) :: a
! Initialize:
a = datetime(2013,1,1,6)
WRITE(*,*)date2num(a) ! 734869.25000000000PURE FUNCTION datetimeRange(d0,d1,t)
! ARGUMENTS:
TYPE(datetime), INTENT(IN) :: d0
TYPE(datetime), INTENT(IN) :: d1
TYPE(timedelta),INTENT(IN) :: tGiven start and end datetime instances d0 and d1,
and time increment as timedelta instance t,
returns an array of datetime instances.
The number of elements is the number of whole time increments
contained between datetimes d0 and d1.
d0 A datetime instance with start time. Will be the first element
of the resulting array.
d1 A datetime instance with end time. Will be the equal to or greater than
the last element of the resulting array.
t A timedelta instance being the time increment for the resulting array.
datetimeRange An array of datetime instances of length
FLOOR((d1-d0)/t)+1
TYPE(datetime) :: a,b
TYPE(timedelta) :: td
TYPE(datetime),DIMENSION(:),ALLOCATABLE :: dtRange
a = datetime(2014,5,1)
b = datetime(2014,5,3)
td = timedelta(days=1)
dtRange = datetimeRange(a,b,td)
! Returns:
!
! dtRange = [datetime(2014,5,1),
! datetime(2014,5,2),
! datetime(2014,5,3)]
a = datetime(2014,5,1)
b = datetime(2014,5,3)
td = timedelta(hours=7)
dtRange = datetimeRange(a,b,td)
! Returns:
!
! dtRange = [datetime(2014,5,1,0),
! datetime(2014,5,1,7),
! datetime(2014,5,1,14),
! datetime(2014,5,1,21),
! datetime(2014,5,2, 4),
! datetime(2014,5,2,11),
! datetime(2014,5,2,18)]PURE ELEMENTAL INTEGER FUNCTION daysInMonth(month,year)
! ARGUMENTS:
INTEGER,INTENT(IN) :: month
INTEGER,INTENT(IN) :: yearReturns the number of days in month for a given month and year.
This function is declared as ELEMENTAL, so it can be called
with scalar or n-dimensional array arguments.
month Integer number of month in year. Valid values are in the range [1-12].
year Integer year.
Returns an integer number of days in requested month and year.
Returns 0 if month is not in valid range.
USE datetime_module,ONLY:daysInMonth
! January on leap year:
WRITE(*,*)daysInMonth(1,2012) ! 31
! February on leap year:
WRITE(*,*)daysInMonth(2,2012) ! 29
! February on non-leap year
WRITE(*,*)daysInMonth(2,2013) ! 28PURE ELEMENTAL INTEGER FUNCTION daysInYear(year)
! ARGUMENTS:
INTEGER,INTENT(IN) :: yearGiven an integer year, returns an integer number of days in that year.
Calls the isLeapYear function.
year An INTEGER scalar or array containing the desired year number(s).
daysInYear An INTEGER scalar or array. Represents the number of days in year.
USE datetime_module,ONLY:daysInYear
! Leap year:
WRITE(*,*)daysInYear(2012) ! 366
! Non-leap year:
WRITE(*,*)daysInYear(2013) ! 365PURE ELEMENTAL LOGICAL FUNCTION isLeapYear(year)
! ARGUMENTS:
INTEGER,INTENT(IN) :: yearReturns a LOGICAL value indicating whether the reqested year is a leap year.
year An INTEGER scalar or array representing year number.
isLeapYear A LOGICAL scalar or array indicating whether a given year is leap year.
USE datetime_module,ONLY:isLeapYear
! Leap year:
WRITE(*,*)isLeapYear(2012) ! .TRUE.
! Non-leap year:
WRITE(*,*)isLeapYear(2013) ! .FALSE.PURE ELEMENTAL TYPE(datetime) FUNCTION num2date(num)
! ARGUMENTS:
REAL(KIND=real_dp),INTENT(IN) :: numGiven the number of days since 0001-01-01 00:00:00 UTC, returns a
correspoding datetime instance.
This function is similar to analogous function in Python (matplotlib.dates.num2date).
num2date is the inverse function of date2num,
so by definition, a == num2date(date2num(a)) evaluates as .TRUE.
for any datetime instance a.
Similarly, b == date2num(num2date(b)) evaluates as .TRUE.
for any variable b of type REAL(KIND=real_dp).
num Number of days since 0001-01-01 00:00:00 UTC.
num2date A datetime instance.
USE datetime_module,ONLY:datetime,num2date
TYPE(datetime) :: a
a = num2date(734869.25d0) ! a becomes datetime(2013,1,1,6,0,0,0)TYPE(datetime) FUNCTION strptime(str,format)
! ARGUMENTS:
CHARACTER(LEN=*),INTENT(IN) :: str
CHARACTER(LEN=*),INTENT(IN) :: formatA wrapper function around c_strptime.
Given a character string str with the format format, returns
an appropriate datetime instance containing that time information.
This function is analogous to Python's datetime.datetime.strptime() function.
Available since version 0.3.0.
str is the character string containing date and time information.
format is the character string containing any combination of regular characters and special format specifiers,
describing the date and time information in str.
Upon successful completion, strptime returns the datetime instance corresponding to the time information contained in str.
Extracting time difference between two time strings using strptime:
USE datetime_module
TYPE(datetime) :: date1,date2
TYPE(timedelta) :: timediff
! Example times in "YYYYMMDD hhmmss" format
CHARACTER(LEN=15) :: str1 = "20130512 091519"
CHARACTER(LEN=15) :: str2 = "20131116 120418"
date1 = strptime(str1,"%Y%m%d %H%M%S")
date2 = strptime(str2,"%Y%m%d %H%M%S")
timediff = date2-date1
WRITE(*,*)timediff
WRITE(*,*)timediff%total_seconds()This example outputs the following:
188 2 48 58 1000
16253339.0000000
This is the same example as in c_strptime but with fewer necessary steps.
PURE ELEMENTAL TYPE(datetime) FUNCTION tm2date(ctime)
! ARGUMENTS:
TYPE(tm_struct),INTENT(IN) :: ctimeGiven a tm_struct instance, returns a corresponding datetime instance. Mostly useful for obtaining a datetime instance after a tm_struct is returned from strptime.
ctime A tm_struct instance.
tm2date A datetime instance.
See example usage for strptime.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent