Extract partial data from datetime using logics/vectorization

Say I have some datetime data:

rowMin =

20×1 datetime array

02-Nov-2020 09:00:00

03-Nov-2020 09:10:00

04-Nov-2020 09:00:00

05-Nov-2020 09:14:59

06-Nov-2020 09:59:59

09-Nov-2020 09:05:00

10-Nov-2020 09:00:00

11-Nov-2020 09:35:00

12-Nov-2020 13:15:00

13-Nov-2020 09:00:00

16-Nov-2020 09:00:00

17-Nov-2020 15:00:00

18-Nov-2020 09:35:00

19-Nov-2020 10:15:00

20-Nov-2020 14:10:00

23-Nov-2020 16:35:00

24-Nov-2020 09:00:00

25-Nov-2020 11:15:00

26-Nov-2020 09:00:00

27-Nov-2020 10:20:00

where I wish to separate different weekdays (always Monday to Friday):

DayNumber=weekday(rowMin);

and then to extract hours and minutes for, respectively, Mondays, Tuesday, ..., Fridays, for subsequent analysis.

How to do this - preferably using logics and vectorization?

My first attempt:

MondayH=zeros(length(rowMin),1); % preallocation

MondayM=zeros(length(rowMin),1); % preallocation

for ii=1:length(rowMin)

if DayNumber(ii,1)==2 % Monday

[MondayH(ii,1),MondayM(ii,1),~]=hms(rowMin(ii,1));

end

end

Bonus info: I would like to utilize dur (duration) on the resulting array so that I afterwards can identify mean HH:mm of the individual days.

Answer 1

[Eric Sofen]

I would extract the time components first and then deal with grouping by day of the week. It's easy to keep this all organized in a table:

d = datetime(2021,1,1) + minutes(randi([0,10000],[100,1])); % cook up some data
d = sort(d);
t = table(d);
t.dow = weekday(d);
t.hour = hour(d);
t.min = minute(d);

% Get the hours for Tuesday (use {} to get results as an array, () as a table)
t{t.dow == 3, "hour"}