Average specific times across a range of dates

I have a large dataset of "minutely" data that has been "cleaned" (there are no duplicates, smaller dt's, or missing minute data)

I have isloated 6 days worth of it in a matlab TimeTable, and want to average together the values at each minute time interval regarless of the day.

So for example if I have:

'6-22-2015 15:04', 5

'6-22-2015 15:05', 2

'6-22-2015 15:06', 3

...

'6-23-2015 15:04', 10

'6-23-2015 15:05', 5

'6-23-2015 15:06', 6

...

'6-24-2015 15:04', 0

'6-24-2015 15:05', 5

'6-24-2015 15:06', 9

I would like it to show:

'15:04', 5

'15:05', 4

'15:06', 6

...

Can Timetables have time values with no associated dates? Should I try and shift the data to all be on the same day?

Answer 1

[Chunru]

% Create the data cell array
x ={
'6-22-2015 15:04',  5
'6-22-2015 15:05',  2
'6-22-2015 15:06', 3
'6-23-2015 15:04',  10
'6-23-2015 15:05',  5
'6-23-2015 15:06', 6
'6-24-2015 15:04',  0
'6-24-2015 15:05',  5
'6-24-2015 15:06', 9};

% Convert it into a time table
t = datetime(x(:,1), 'InputFormat', 'mm-dd-yyyy HH:mm');
v = cell2mat(x(:, 2));
tt =timetable(t, v);

% For example, find the data has same hour and minute as '15:04'
str = '15:04'
idx = minute(tt.t) == minute(str) & hour(tt.t) == hour(str)
av = mean(tt.v(idx));

% The result:
str, av