Converting dates in time series

Dear All, I have a time seris that containts the time vector as :

27-Feb-2007 12:00:00 58.199 -136.64 -0.0048352

28-Feb-2007 12:00:00 58.199 -136.64 -0.0048265

01-Mar-2007 12:00:00 58.199 -136.64 -0.0048853

02-Mar-2007 12:00:00 58.199 -136.64 -0.0050604

03-Mar-2007 12:00:00 58.199 -136.64 -0.0051769

04-Mar-2007 12:00:00 58.199 -136.64 -0.0052122

I need to change the time format of the time column (first column) to decimal year format. I need the time format something like 2007.0123. please help me

Answer 1

[Davide Masiello]

Take a look at this

date = '27-Feb-2007 12:00:00';
str = datestr(datenum(date),'yyyy.dd.mm')

str =

    '2007.27.02'
    

And for more info you can take a look at the following webpages

https://uk.mathworks.com/help/matlab/ref/datenum.html

https://uk.mathworks.com/help/matlab/ref/datestr.html

Answer 2

[Star Strider]

One approach —

M = {'27-Feb-2007 12:00:00'                58.199                           -136.64                 -0.0048352 
     '28-Feb-2007 12:00:00'                58.199                           -136.64                 -0.0048265 
     '01-Mar-2007 12:00:00'                58.199                           -136.64                 -0.0048853 
     '02-Mar-2007 12:00:00'                58.199                           -136.64                 -0.0050604 
     '03-Mar-2007 12:00:00'                58.199                           -136.64                 -0.0051769 
     '04-Mar-2007 12:00:00'                58.199                           -136.64                 -0.0052122};


format long 
T1 = cell2table(M)

T1 = 6×4 table
               M1                 M2        M3           M4    
    ________________________    ______    _______    __________

    {'27-Feb-2007 12:00:00'}    58.199    -136.64    -0.0048352
    {'28-Feb-2007 12:00:00'}    58.199    -136.64    -0.0048265
    {'01-Mar-2007 12:00:00'}    58.199    -136.64    -0.0048853
    {'02-Mar-2007 12:00:00'}    58.199    -136.64    -0.0050604
    {'03-Mar-2007 12:00:00'}    58.199    -136.64    -0.0051769
    {'04-Mar-2007 12:00:00'}    58.199    -136.64    -0.0052122

ddoy = day(datetime(T1.M1),'dayofyear')/day(datetime(year(T1.M1),12,31),'dayofyear');
% T1.M1 = year(datetime(T1.M1)) + ddoy(:,1)                           % Not Rounded

T1.M1 = year(datetime(T1.M1)) + round(ddoy(:,1), 4)                 % Round To 4 Decimal Places

T1 = 6×4 table
       M1          M2        M3           M4    
    _________    ______    _______    __________

    2007.1589    58.199    -136.64    -0.0048352
    2007.1616    58.199    -136.64    -0.0048265
    2007.1644    58.199    -136.64    -0.0048853
    2007.1671    58.199    -136.64    -0.0050604
    2007.1699    58.199    -136.64    -0.0051769
    2007.1726    58.199    -136.64    -0.0052122

I hope that there are easier ways to do this using datetime functions. I looked, however there doesn’t appear to be a format option for this.

EDIT — (7 Mar 2022 at 14:20)

Added round call to ‘ddoy’ display.

.

[Chris Martin]

does it take account the leap years ?

[Star Strider]

The datetime function (and related functions) accounts for leap years, leap seconds, and every other known time variant.

Answer 3

[Peter Perkins]

Here's my solution:

>> tt = timetable(rand(6,1),rand(6,1),'RowTimes',datetime(2007,2,27:32,12,0,0))
tt =
  6×2 timetable
            Time             Var1       Var2  
    ____________________    _______    _______
    27-Feb-2007 12:00:00    0.14929    0.92926
    28-Feb-2007 12:00:00    0.25751    0.34998
    01-Mar-2007 12:00:00    0.84072     0.1966
    02-Mar-2007 12:00:00    0.25428    0.25108
    03-Mar-2007 12:00:00    0.81428    0.61604
    04-Mar-2007 12:00:00    0.24352    0.47329
>> lengthOfYear = @(t) caldays(between(dateshift(t,'start','year','current'),dateshift(t,'start','year','next'),'days'))
lengthOfYear =
  function_handle with value:
    @(t)caldays(between(dateshift(t,'start','year','current'),dateshift(t,'start','year','next'),'days'))
>> 
>> tt.DecimalYear = year(tt.Time) + day(tt.Time,'dayofyear')./lengthOfYear(tt.Time)
tt =
  6×3 timetable
            Time             Var1       Var2      DecimalYear
    ____________________    _______    _______    ___________
    27-Feb-2007 12:00:00    0.14929    0.92926      2007.2   
    28-Feb-2007 12:00:00    0.25751    0.34998      2007.2   
    01-Mar-2007 12:00:00    0.84072     0.1966      2007.2   
    02-Mar-2007 12:00:00    0.25428    0.25108      2007.2   
    03-Mar-2007 12:00:00    0.81428    0.61604      2007.2   
    04-Mar-2007 12:00:00    0.24352    0.47329      2007.2   

SS, I think this is just about what you did, except for putting the length of year calculation in a function.

Chris, unless you need to export these values, I'd think twice about creating them. It's kind of a terrible time representation in which you get poor precision, and even whole days are subject to round-off errors. SS, that kind of answers why there is not a simpler way to do this, but maybe this is a more common format that I'm aware of.