How can I convert datetime into double?

I have a data which is 60 x 4320 datetime, each cell is 2019/01/01 00:00:00 and want to convert this into 60 x 4320 double.

I tried to use datevec(data) but it resulted 259200 x 6 double, one row one character, like

2019 01 01 00 00 00
2019 01 01 00 10 00
2019 01 01 00 20 00

[Stephen23]

datenum does what you ask for... but probably is not what you need: what do you want to do with double that you don't think is possible with datetime?

[MatLab Code N]

it is not possible to merge datetime with other double variables

[Steven Lord]

You cannot have a double array where some of the elements are datetime values, that is true. But if you have date- and/or time-stamped data, I suggest using a timetable instead.

T = datetime('today');
randomDays = T + days(randi([-7 7], 10, 1));
x = (1:10).';
myTimedData = timetable(randomDays, x)

myTimedData = 10×1 timetable
    randomDays     x 
    ___________    __

    15-Nov-2021     1
    11-Nov-2021     2
    16-Nov-2021     3
    18-Nov-2021     4
    13-Nov-2021     5
    11-Nov-2021     6
    22-Nov-2021     7
    13-Nov-2021     8
    15-Nov-2021     9
    11-Nov-2021    10

Answer 1

[galaxy]

You can use datenum to convert each your data to double.

Example:

>> t = datetime('now')

t = 

  datetime

   2019/12/12 09:43:30

>> DateNumber = datenum(t)

DateNumber =

   7.3777e+05

After that, if you want to convert back:

   >> datetime(DateNumber,'ConvertFrom','datenum')

ans = 

  datetime

   2019/12/12 09:43:30

hope it helps

Answer 2

[Eric Sofen]

What do you want the doubles to represent?

Datenum returns a double represents days since 0 CE.

The posixtime function would give you a double representing seconds since the Unix epoch (00:00:00 1-Jan-1970 UTC).

The juliandate function gives a double representing days since...

The trouble with using double to represent dates and times is that there isn't one definition (there are also precision-related issues over long spans of time).

I'm curious if you'd be willing to share what you want to do with the doubles, and why you can't do it using datetime?

Answer 3

[Tianchu Lu]

Hi,

I have got the same problem, and I think I have managed to solve it.

This function does exactly what it is. Although this is for a cell array, just change {} brace bracket to () round bracket if your data type is not a cell type.

But please let me know if you are going to use it.

Thanks

Tian

function datetime_double_matrix = convertDatetimeCellArray(datetime_array)

% Convert a 1xN cell array of datetime strings to a 6xN double matrix

% in the order of [YYYY, MM, DD, HH, MN, SS]

% Initialize a matrix to store the date vectors

num_entries = length(datetime_array);

datetime_double_matrix = zeros(6, num_entries);

% Loop through the cell array and convert each datetime to a date vector

for i = 1:num_entries

% Convert the datetime string to a date vector

date_vector = datevec(datetime_array{i});

% Transpose the date vector and store it in the matrix

datetime_double_matrix(:, i) = date_vector';

end

end

[Adam Danz]

You could also just use datevec(datetimeStrings)'

dtstr = {'12/10/2020 10:22:03','1/24/2021 22:00:01','9/9/1999'};
convertDatetimeCellArray(dtstr)

ans = 6x3

2020 2021 1999 12 1 9 10 24 9 10 22 0 22 0 0 3 1 0

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

datevec(dtstr).'

ans = 6x3

2020 2021 1999 12 1 9 10 24 9 10 22 0 22 0 0 3 1 0

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

function datetime_double_matrix = convertDatetimeCellArray(datetime_array)
    % Convert a 1xN cell array of datetime strings to a 6xN double matrix
    % in the order of [YYYY, MM, DD, HH, MN, SS]
    % Initialize a matrix to store the date vectors
    num_entries = length(datetime_array);
    datetime_double_matrix = zeros(6, num_entries);
    % Loop through the cell array and convert each datetime to a date vector
    for i = 1:num_entries
        % Convert the datetime string to a date vector
        date_vector = datevec(datetime_array{i});
        
        % Transpose the date vector and store it in the matrix
        datetime_double_matrix(:, i) = date_vector';
    end
end