how to horizontally concatenate columns of different types

Hello, In my workspace I have a variable Z with strings (5082 x 1 cell), and a variable Y with numbers (5082 x 4360 single).

I would like to concatenate these two variables horizontally in order to obtain a matrix or something that I can then export into Excel/text file, with 5082 rows and 4361 columns.

I also need the column with strings to be the first column of the matrix/array/etc.

I have tried with: matrix={[char(Z),num2str(Y)]}; but it doesn't do what I want. It gives me a 1 x 1 cell.

I have also tried: mixed_array = cat(1, Z, Y); and again it doesn't do what I want. It results in a 5083 x 1 cell, with the 5082 rows from the variable Z and the last row (5083th) with the entire Y variable (5082 x 4360 single).

If I do: mixed_array = cat(2, Z, Y); it gives me this Error: "Error using cat Dimensions of matrices being concatenated are not consistent."

Any suggestions about how to do it?

Many thanks!

Answer 1

[the cyclist]

I think your best bet is for the final object to be a cell array, so try this:

Z = cell(5082,1);
Y = single(rand(5082,1));

% Put each numeric into a cell, then concatenate:
M = [Z,num2cell(Y)];

Also, I like this cellwrite function from the File Exchange for writing cell arrays to CSV files. (I am on a Mac, so the MATLAB tools are somewhat crippled.)

[dederi]

@the cyclist, Thanks for your answer! I tried to use your script: M = [Z,num2cell(Y)]; but my computer crashes for more than 3 hours, and after this time, still no output. Do you think that the problem is the memory of my computer? Or is normal that this formula requires a lot of time to produce an output? Many thanks.

[the cyclist]

The exact code in my example took 0.01 seconds to run on my computer.

I suggest running your code on a very small subset of your data, and see if it returns quickly. Then maybe you could profile your code to see what is so slow.