Question

Hi I have a 3 dimensional matrix that I am trying to convert the rows to columns and then stack all the columns to generate a vector that will have the three dimensions.

therefore, row 1 becomes column 1, row 2 becomes column 2 etc. And then, column 2 goes under column 1, column 3 under column 1 and 2. I also need to keep the dimensions the same

PV_power_output(:,:,K) below is a 365 x 24 x 27 matrix, which should become a 8760×1 vector following the positioning sequence mentioned above

I am struggling with the coding of this

So far I have

PV_power_output(:,:,K) = real((Vmpp_(:,:,K).*Impp_(:,:,K))*inverter_efficiency)/1000; 

 permdims = 1 : ndims(PV_power_output); 
out = reshape(permute(PV_power_output, permdims),8760,1,[]);

However, on checking the elements in different positions, the above is incorrect and so any help would be great!

Answer 1

Perhaps your permute is wrong, because as it stands it’s not doing anything:

permute(PV_power_output, 1:3) == PV_power_output

maybe you need:

permute(PV_power_output, [2 1 3])

I general I often find reshape difficult to follow (especially when you go back to the code three weeks later) so if performance is not critical then you could consider using a loop here instead for your 3d matrix.

You might think about a cell array instead of 3d matrix, and then your reshape is much more straight forward:

PV_power_output{K} = real((Vmpp_(:,:,K).*Impp_(:,:,K))*inverter_efficiency)/1000;

vector = reshape(PV_power_output{K},8760,1);

---

Update:

Perhaps try reshaping to 2D:

out = reshape(permute(PV_power_output, [2 1 3]), 8760,27);

And then you can access the Kth column vector as:

vector_K = out(:,K)