finding max for each 3 columns

Hi All,

I have a dataset something similar to following

   Name     Location     Nickname     saleCost
   rest1    east         Reast1       [5, 7, 9 , 2, 5, 1, 9, 11, 1]
   rest4    south        Rsouth4      [12, 9, 2, 3, 5, 10, 11, 4, 4]
   rest3    east         Reast3       [20, 2, 2, 3, 20, 30, 1, 3, 7]
   rest9    west         Rwest9       [1, 9, 3, 33, 22, 10, 6, 6, 8]

Note that saleCost is a dataset as well and as number of columns which are multiple of 3.

I want to find max for every 3 columns so my resultant dataset is as follow

     Name     Location     Nickname     saleCostMax
     rest1    east         Reast1       [9, 9, 9, 5, 9, 11, 11, 11, 1]
     rest4    south        Rsouth4      [12, 9, 5, 10, 11, 11, 11, 4, 4]
     rest3    east         Reast3       [20, 3, 20, 30, 30, 30, 7, 7, 7]
     rest9    west         Rwest9       [9, 33, 33, 33, 22, 10, 8, 8, 8]

ANy other way so that I do not have to use for loop? WHat are my options?

Thanks

[Geoff Hayes]

Neesha - is your data set a cell array or something else? Why don't you want to use a for loop?

[Neesha]

isn't for loop time consuming and not the best approach? I will have 100s of lines

Answer 1

[Image Analyst]

If you have the Image Processing Toolbox, you can go down row by row and call imdilate() to get a moving max:

thisRow = [5, 7, 9 , 2, 5, 1, 9, 11, 1]
newRow = imdilate(thisRow, [1,1,1])
% Crop off the first element and tack on the last element
newRow = [newRow(2:end), thisRow(end)]

In the command window:

thisRow =
     5     7     9     2     5     1     9    11     1
newRow =
     7     9     9     9     5     9    11    11    11
newRow =
     9     9     9     5     9    11    11    11     1

[Neesha]

seems it worked in this example for me. :) if i have to take max of every 6 instead of 3, do i have keep changing newRow? How?

[Image Analyst]

The kernel in imdilate() defines what elements in a sliding window to look at. If you wanted to look at 6 use [1,1,1,1,1,1]. Use a zero if you want to ignore an element.

[Neesha]

ok makes sense. I tried and this is what i get. It is not correct output

thisRow = [5, 7, 9 , 2, 5, 1, 9, 11, 1]
newRow = imdilate(thisRow, [1,1,1,1,1,1])
% Crop off the first element and tack on the last element
newRow = [newRow(2:end), thisRow(end)]

thisRow =

     5     7     9     2     5     1     9    11     1

newRow =

     9     9     9     9     9    11    11    11    11

newRow =

     9     9     9     9    11    11    11    11     1

[Image Analyst]

What would you want? A sliding window of 6 only fits completely in there in 4 spots. Do you want 4 output numbers, or do you want the window to slide off on both ends?