Apply function on specific rows

Hi Everyone,

I am kind of new to MATLAB and have a problem. I used to use, S-Plus and they have a function called "tapply" and I think I am looking for something similar.

Here is the situation, I have time-series database, with a column representing month/year. I would like to apply "mean" function to those rows which have same month/year. For example, if there are 20 unique month/year values, i will have 20 different "mean" each representing, an average of that particular month/year.

Looking forward to responses. Thank you in advance.

[Fangjun Jiang]

It should be quite easy if you provide some example data.

Answer 1

[Walter Roberson]

You could use a "for" loop in the general case.

If you are computing a value for each row, you could use

TheValues = arrayfun(@(K) YourFun(YourVariable(K,:)), [list of rows])

For simple functions such as mean, and simple arrays, you can use

TheValues = mean(YourVariable([list of rows],:),2);

The ,2 part tells it to take the mean along the second dimension (rows)

Answer 2

[KKR]

Thank you Walter, but I am little confused what is K? And what do you mean by list of rows?

Fangjun Jiang,

I have attached a sample data base. Please see if it helps.

Kaushal

date  tr
1/1/2003  1.880878
1/1/2003  -19.07757
1/1/2003  -20.79946
1/1/2003  -10.12233
1/1/2003  -4.862288
4/1/2003  26.86057
4/1/2003  21.90327
4/1/2003  17.20895
4/1/2003  40.25018

[Walter Roberson]

The K is the arglist in the anonymous function; see http://www.mathworks.com/help/techdoc/matlab_prog/f4-70115.html

A list of rows would be like [1 2 3 4 5] or [6 7 8 9] or 1:5 or 6:9 to apply to the designated rows. Your original question did not indicate that your data would always be consecutive so I was thinking that you might have to specify (e.g.) [1 2 6 8 10 11 15]

I do not know how to retrieve a particular row from a database. If the above data were presented in the form of a 9 x 2 cell array, then I would use code such as

mean(vertcat(data{1:6,2}))

To go beyond this, see unique() and ismember(), and look in the MATLAB File Exchange to see if John d'Errico's "consolidator" program is available.