How to create a loop which nan means every datapoint of three components

I have 3 datasets at 11564597 data points each - my calculation is U = nanmean(sqrt((u).^2+(v).^2+(w).^2)); to get the mean flow of the 3 sinusoidal curves. To get an array that has 11564597 points to plot on top of the 3 components as the mean flow of the three compontents.

My code so far is:

U= NaN*ones(length(velocity_x_nortek.smooth));
for i = 1:length(velocity_x_nortek.smooth)
    U(i)= nanmean(sqrt((velocity_x_nortek.smooth).^2+(velocity_y_nortek.smooth).^2+(velocity_z_nortek.smooth).^2);
end

Answer 1

[ADragon]

Hi Rebecca, you do not need the for loop since the equation can operate on arrays. Just use:

U = nanmean(sqrt((velocity_x_nortek.smooth).^2+(velocity_y_nortek.smooth).^2+(velocity_z_nortek.smooth).^2);
plot([velocity_x_nortek.smooth velocity_y_nortek.smooth velocity_z_nortek.smooth U])

Assuming you have column vectors for plotting.

AD

[Rebecca Ellis]

U Just give me one number, rather than 11564597 data point?

[ADragon]

Do you have row vector or column vector? You may need to add a second input argument into nanmean function or transpose your vectors. Personally I would use the below code if you have row vectors.

U = nanmean(sqrt((velocity_x_nortek.smooth').^2+(velocity_y_nortek.smooth').^2+(velocity_z_nortek.smooth').^2);

[Rebecca Ellis]

they are within different structures in a column as attached. When I use that code I get one point only. Do I need to turn them into column vectors and how would I do this? Thank you for your help very very appreciated.

[ADragon]

Your variables are column vectors already. Maybe I misunderstood your equation. To get the mean by row of velocity_x_nortek.smooth, velocity_y_nortek.smooth, and velocity_z_nortek.smooth, use:

U = nanmean([velocity_x_nortek.smooth velocity_y_nortek.smooth velocity_z_nortek.smooth],2)

[Rebecca Ellis]

The calculation is

   U = nanmean(sqrt((u).^2+(v).^2+(w).^2))
is correct - 
I just need to loop it that it does the calculation for every 11564597 points -

U = zeros(length(timevec), 1);

for i = 1:length(velocity_x_nortek.smooth)
    U(i)= nanmean(sqrt((velocity_x_nortek.smooth).^2+(velocity_y_nortek.smooth).^2+(velocity_z_nortek.smooth).^2));
end

The loop I have doesn't work ?

[ADragon]

OK. You can add an index to your vector.

for i = 1:length(velocity_x_nortek.smooth)
    U(i)= nanmean(sqrt((velocity_x_nortek.smooth(i)).^2+(velocity_y_nortek.smooth(i)).^2+(velocity_z_nortek.smooth(i)).^2));
end

But you will be taking the mean of a scalar value. So I am confused as to what you are trying to do...

If you are trying to calculate RMS then switch mean and sqrt.

for i = 1:length(velocity_x_nortek.smooth)
      U(i)= sqrt(nanmean((velocity_x_nortek.smooth(i)).^2+(velocity_y_nortek.smooth(i)).^2+(velocity_z_nortek.smooth(i)).^2));
  end