Extract time (hour+minute+second) from datetime vector

I have an array that has datetime in one column, and a rate in the second column. i want to make a plot that just has on the x axis, a 24 hour day, and I want to see how the data varies over the course of the day. My dataset is a month long - so I'd like to plot all of that data and visualize it by time of day.

i don't know how to extract just the hour+minute+second data in ONE column. I know I can get hour separately, minute separately, etc. but I just want a column that says 01:30:22 for example, for each datetime value I have. I've looked at the function datetime but I'm pretty sure that gives me day as well? I JUST want time in the output

For reference, my table (BNFrates) has datetime in this format: 738771.683333333

Thank you!

[Dyuman Joshi]

Could you please attach the data file? Use the paperclip button to attach.

[Ariana]

here is the data file!

Answer 1

[Ariana]

hi everyone,thank you to everyone who answered! I figured it out, and this is the code I used:

If datetimeBNF is the column with the original datetimes:

formattedDatetimes = cell(size(datetimeBNF));
for i = 1:numel(datetimeBNF)
    formattedDatetimes{i} = datestr(datetimeBNF(i), 'HH:MM:SS');
end

Answer 2

[Steven Lord]

If you can use a datetime array instead of serial date numbers (or convert the serial date numbers to datetime):

rightnow = datetime('now')

rightnow = datetime
   25-Oct-2023 18:07:00

use the timeofday function.

[timeSinceMidnight, midnight] = timeofday(rightnow)

timeSinceMidnight = duration
   18:07:00

midnight = datetime
   25-Oct-2023

To convert, first let's get the serial date number (you wouldn't need to do this, as you already have it. I need to do it so I can show you the results.)

format longg
serialDateNumber = datenum(rightnow)

serialDateNumber = 
          739184.754871983

Then call datetime with an option:

rightnow2 = datetime(serialDateNumber, 'ConvertFrom', 'datenum')

rightnow2 = datetime
   25-Oct-2023 18:07:00

[Ariana]

Thank you Steven, this is also very helpful!!

Answer 3

[Dyuman Joshi]

Here's an approach -

in = load('example_array.mat')

in = struct with fields:
    BNFratesall_NaN1: [21×2 double]

vec = in.BNFratesall_NaN1;

%Convert the dates to datetime()
dt = datetime(vec(:,1), 'ConvertFrom', 'datenum')

dt = 21×1 datetime array
   07-Sep-2022 16:22:00
   07-Sep-2022 16:23:59
   07-Sep-2022 16:25:59
   07-Sep-2022 16:27:59
   07-Sep-2022 16:30:00
   07-Sep-2022 16:32:00
   07-Sep-2022 16:34:00
   07-Sep-2022 16:36:00
   07-Sep-2022 16:38:00
   07-Sep-2022 16:40:00
   07-Sep-2022 16:41:59
   07-Sep-2022 16:43:59
   07-Sep-2022 16:45:59
   07-Sep-2022 16:48:00
   07-Sep-2022 16:50:00
   07-Sep-2022 16:52:00
   07-Sep-2022 16:54:00
   07-Sep-2022 16:56:00
   07-Sep-2022 16:57:59
   07-Sep-2022 16:59:59
   07-Sep-2022 17:01:59

%Get the hour, minute and second values
[h,m,s] = hms(dt);

%Get the output as a duration() array
out = duration(h,m,s)

out = 21×1 duration array
   16:22:00
   16:23:59
   16:25:59
   16:27:59
   16:30:00
   16:32:00
   16:34:00
   16:36:00
   16:38:00
   16:40:00
   16:41:59
   16:43:59
   16:45:59
   16:48:00
   16:50:00
   16:52:00
   16:54:00
   16:56:00
   16:57:59
   16:59:59
   17:01:59