Computing average times from an array of datetime using only selected rows

Hi there, I have an array of datetime variables, and an array with a number in each row which corresponds to how I want to 'group' the datetime array. For example, my datetime array might look like:

>>exact_times =

8×1 datetime array

   21-Jan-2016 11:24:17
   21-Mar-2018 11:24:17
   21-Mar-2018 11:24:17
   21-Mar-2018 11:24:17
   21-Mar-2018 11:25:36
   21-Mar-2018 11:25:36
   21-Mar-2018 11:25:47
   21-Mar-2018 11:25:47

and my subs array might look like

>> subs

subs =

     2
     1
     1
     2
     1
     4
     4
     4
     1
     3

I want to average all the rows in exact_times which correspond to a given number in subs, ie all the rows which correspond to 1 in subs (this is rows 2,3,5,9) and then do the same for 2 in subs, etc.

If I could use accumarray I would use something like: average_time = accumarray(subs, exact_times, [], @mean); but I can't use that on datetime variables. I can't change the ordering of anything in either array for other reasons. Can anyone help me out?

Thanks in advance!

Answer 1

[supernoob]

I figured it out: simply convert to datenum and then back:

average_time = accumarray(subs, datenum(exact_times), [], @mean); avg_data.average_time = datetime(average_time, 'ConvertFrom', 'datenum');