First calculating then looping with multiple figures

Hi all

I am trying to combine some matrices, all with the same dimensions but different names. More specifically the quantities I am trying to assess have the same name of variable but a different timestep indicating their recorded time (e.g. Windv_y_20040501_180000)

I usually have used, when i first time started to analyse them, which give me a moving figure in matlab

[a1 a2 a3]=size(Windv_y_20040501_180000);

v10_profile=reshape(Windv_y_20040501_180000,a1,a2*a3);

for a=1:a3
     imagesc(V10(:,:,a))
    title('wind profile m/s V_1_0'),colorbar
     drawnow;
     pause(0.2);
end

Although this does not satisfy the next thing i want to achieve which is to combine two quantities and then have a figure(subplot) with the new combined re-calculated matrices in a separate plot

The files i normally use have this kind of structure, a constant name of variable but different endings.

Windv_y_20040501_180000
Windv_y_20040501_210000

Windv_x_20040501_180000
Windv_x_20040501_210000

Assuming that i want to automate a process for example, adding the corresponding matrices and then plot them

m1= Windv_x_20040501_180000 + Windv_y_20040501_180000
m2= Windv_x_20040501_210000 + Windv_y_20040501_210000

figure 
subplot(2,1,1),image(m1)
subplot(2,1,2),image(m2)

Any suggestion would be welcome

thank you

[George]

in the calculation i forgot to insert the (.) operator sorry

m1= Windv_x_20040501_180000.+ Windv_y_20040501_180000
m2= Windv_x_20040501_210000. + Windv_y_20040501_210000

[Geoff Hayes]

George - how many of these variable pairs do you have? Do you read them from file to create the two different (x,y) variables, or do something else?

Why not just read all of the data into a cell array, and then you can avoid having to be concerned with the different variable names. This would simplify the code considerably.

[George]

The number of files varies (from 4to 26, meaning two from each quantity), I would like to avoid merging them into a single variable, because for post processing purposes, I have to keep the m1 , m2 ...mn for future usage.

As a first step i want to somehow iterate the appropriate varnames into a formulation I have constructed to get m1 (the addition here is just an example).

[Jon Boerner]

I am still not quite sure what your whole workflow is, but going on what George suggested, you could keep the variables in a larger structure so that their names are maintained. For example:

mydata.Windv_x_20040501_180000 = Windv_x_20040501_180000;
mydata.Windv_y_20040501_180000 = Windv_y_20040501_180000;

Then you could iterate over the variables using the fieldnames function:

>> fields = fieldnames(mydata)

ans = 

      'Windv_x_20040501_180000'
      'Windv_y_20040501_180000'

And then using a loop:

for i=1:length(fields)
    m = mydata.(fields{i});
    % Plotting code...
end

You could do some more complicated logic to pull out fields with the same timestep and add them together or something like that using the field names as well.

Answer 1

[Geoff Hayes]

You could create a cell array of n elements and so still keep the data sets distinct. If you are reading from n files then it would be something like

 windVData = cell(n,2);

 for k=1:n
    % read x and y data from file
    % some code here

    % save to cell array
    windVData(k,1) = xData;
    windVData(k,2) = yData;
 end

You could even save the file name in a third column of the array (if you wish).

Then you could iterate over this cell array and produce your subplots

 figure;
 n = size(windVData,1);

 for k=1:n
     % create a single column of n subplots
     h = subplot(n,1,k);

     % add the two vectors (assumed to be of the same dimension)
     m1= windVData{k,1} + windVData{k,2};

     % display the data
     image(m1,'Parent',h);
 end

Not really clear if you wish to use image or plot instead.

[George]

The data are generated by a code I have written in through a fortran process algorithm and are directly save to a .mat file (thus they dont occupy space)

Each matrix has 300x300 dimensions or more , so i think the image plot is more accurate for me.

I didt't got the part you explain the m1 generation, does windVData{k,1} stands for the first column?

Is for

i=1:20041231_180000
m{i}=Windv_y_{i}+Windv_x_{i}
end

And for the plotting, what would you suggest?

[Geoff Hayes]

Right - the two columns in the cell array correspond to the x and y data for a specific time step. The above just gives a general idea of what you can do (since you haven't shown how you load/read the different variables or time steps).

Though I'm not sure what you are hinting at with your assignment of i, as 20041231_180000 cannot be used as an index into m or suffixed onto Windv_y_.

As for plotting, is your data the wind velocity in the x direction and wind velocity in the y direction? What does the 300x300 mean? What does image show you when you add the two together? Since you haven't described what this data represents, I can't provide any (good) suggestions on how it should be displayed.

[George]

By combining the components I get the overall wind resource, each component basically is 300x300 matrix (or more) and represents a gridded mesh, with the each point of the having a assigned U, or V value, thus the existence of the tableS. Image basically is used, either that or contour to show the mesh itself