Find duplicate entries and average associated values

Hi all,

I have a 180 x 2 array of data and in the first column there are some repeated values (multiple data points at the same location). With those repeated values I'd like to average the associated data in column 2. Any help with this would be appreciated greatly.

Example array: [1 10; 2 20; 3 30; 4 40; 4 50; 4 60; 5 70; 6 80; 7 90; 7 100; 8 110]

I'd like to create a new array that looks like this: [1 10; 2 20; 3 30; 4 50; 5 70; 6 80; 7 95; 7 100; 8 110]

Thanks a ton for any help.

-Chris

[Cedric]

Are elements from the first column positive integers?

Answer 1

[José-Luis]

 array =  [1 10; 2 20; 3 30; 4 40; 4 50; 4 60; 5 70; 6 80; 7 90; 7 100; 8 110];
 [C,ia,idx] = unique(array(:,1),'stable');
 val = accumarray(idx,array(:,2),[],@mean); 
 your_mat = [C val]

Answer 2

[Cedric]

If elements from the first column are "continuous" positive integers, here is a way to go:

 >> data = [1 10; 2 20; 3 30; 4 40; 4 50; 4 60; 5 70; 6 80; 7 90; 7 100; 8 110] ;
 >> newData = [unique( data(:,1) ), ...
               accumarray( data(:,1), data(:,2), [], @mean )]
 newData =
     1    10
     2    20
     3    30
     4    50
     5    70
     6    80
     7    95
     8   110

If you don't have "continuity" (i.e. there are gaps), here is a more general solution (I removed row 5 70 from previous data):

 >> data = [1 10; 2 20; 3 30; 4 40; 4 50; 4 60; 6 80; 7 90; 7 100; 8 110] ;
 >> uc1 = unique( data(:,1) ) ;
 >> mc2 = accumarray( data(:,1), data(:,2), [], @mean ) ;
 >> newData = [uc1, mc2(uc1)]
 newData =
     1    10
     2    20
     3    30
     4    50
     6    80
     7    95
     8   110

Answer 3

[derGerät]

Hey all, I have kind of the same problem but my data ar not in one array but two arrays. Is it possible to adept this solution to make it work with two arrays?

Sorry if thats a stupid question but I'm a beginner...

Answer 4

[Chris]

This is great, but the first column elements are actually not integers. They are continuous, but they have precision up to three decimal points.

Thanks for the quick reply.

Answer 5

[Chris]

Incredible, thanks so much José-Luis and Cedric. I've just learned so many new things from these few lines of code.

[José-Luis]

No worries.

Please accept an answer if it helped you.

[Cedric]

My pleasure, note that I assumed integers because, as far as I am concerned, it is rarely appropriate to use UNIQUE on floating points, for the same reason it is rarely appropriate to test equality between floats.

Answer 6

[Kalpana Patel]

how to filter the repeated datain a data series maintaining the length. I have two column, one is for latitude and second is for ion density, with the use of unique code length is maot maintain. reason i know. so just want single latitude valua and data corresponding to it. total data values are 4892. pls help..