Uneven time series averaging

Need to get hourly averages of data which have uneven frequency over a given hour. Id is the 'sample' which is repeated over and over. Here is example data:

 id  hour  data
  1  17  376.2
  2  17  376.1
  3  17  375.9
  4  17  375.7
  5  17  375.9
  1  17  375.7
  2  18  375.9
  3  18  376.1
  4  18  376.3
  5  18  376.6
  1  18  377.0
  2  18  377.2
  3  19  377.5
  4  19  377.7
  5  19  378.0
  1  19  378.1
  2  19  378.4
  3  19  378.6
  4  19  378.7
  5  20  378.8

I want to arrive at something this this in which the hours having 2 samples get a mean value:

hour  id1  id2  id3  id4  id5
17  376.0  376.1  375.9  375.7  375.9
18  377.0  376.5  376.1  376.3  376.6
19  378.1  378.4  378.0  378.2  378.0
20  379.1  379.1  379.3  378.8  379.3

I have unique value index for HH and id having already used accumarray to collect the data as shown. Is there a way to use the 'aggregation function' of unstack() to do this?

newdata = unstack(data, {'data1'}, 'id');

I can't figure out the syntax based on mathworks help. Thanks....

[Colin Edgar]

Got a blank answer..... thanks in advance for any help!

[Colin Edgar]

dataout = unstack(datin, {'data'}, 'id', 'AggregationFunction', @mean);

Worked out the syntax for using unstack() in this way but doesn't solve my problem.

[Colin Edgar]

I solved my problem using accumarray(). Will post code soon.

Answer 1

[Colin Edgar]

http://www.mathworks.com/matlabcentral/answers/252435-group-mean-for-repeated-samples

I adapted this approach to work here. The key is correct use of unique() to get the index that identifies the groups you are working with.