Average data in an array based on corresponding time values

Hi there:

Let's say I have an array t = [3,1,5,6,2,1,3] and a corresponding data array d = [352,59592,5783,27,6849,264,284]. I want to average all of the values in d that correspond to the values in t. For example, when t = 1, I want to average 59592 and 264. Then, I want a new array t1 = [1,2,3,5,6] to correspond to the averages that were computed for each time and stored in a new array, d1. This is just a smaller example for a larger dataset.

This is what I'm trying to do right now:

for hour_to_test = 0:23
    for i=1:length(t)
        if t(i) == hour_to_test
            temp = [temp, d(i)];
            n = mean(temp);
            d1 = [d1,n];
        end
        temp = [ ];
    end
end

But this doesn't seem to be averaging all of the values for each hour. I am getting an array out of this loop that is the same length at d, not a length of 24. How do I go about fixing this?

Answer 1

[Star Strider]

Try this:

t = [3,1,5,6,2,1,3];
d = [352,59592,5783,27,6849,264,284];
[tu,~,ix] = unique(t);
dmeans = accumarray(ix, d, [], @mean);
d1 = [tu(:), dmeans]

d1 =
           1       29928
           2        6849
           3         318
           5        5783
           6          27

[Megan Henriksen]

Thank you so much!!

[Star Strider]

As always, my pleasure!

Answer 2

[Peter Perkins]

Star Strider's solution is old school. Nothing wrong with that, it gets the job done, but many people find accumarray hard to fathom. Here's another approach that uses timetables, which you may find makes other things easier:

>> t = [3;1;5;6;2;1;3];
>> d = [352;59592;5783;27;6849;264;284];
>> tt = timetable(d,'RowTimes',seconds(t))
tt =
  7×1 timetable
    Time       d  
    _____    _____
    3 sec      352
    1 sec    59592
    5 sec     5783
    6 sec       27
    2 sec     6849
    1 sec      264
    3 sec      284
>> ttAvg = retime(tt,'secondly','mean') % one way
ttAvg =
  6×1 timetable
    Time       d  
    _____    _____
    1 sec    29928
    2 sec     6849
    3 sec      318
    4 sec      NaN
    5 sec     5783
    6 sec       27
>> ttAvg = varfun(@mean,tt,'GroupingVariable','Time') % another way
ttAvg =
  5×2 timetable
    Time     GroupCount    mean_d
    _____    __________    ______
    1 sec        2         29928 
    2 sec        1          6849 
    3 sec        2           318 
    5 sec        1          5783 
    6 sec        1            27 

[Andrei Bobrov]

+1!!!