help with dynamic variable names set with 'for' incrementer

Hi,

I am trying to split up a large numerical dataset into individual days. What I have that works and what I have that failed is;

function [FxStOut, FxStIndex] = FxSt(FxStIn)
%Struct for Forex

UnqDate = unique(FxStIn(:,1));
UnqL = length(UnqDate);

for ii = 1:UnqL
    indx = FxStIn(:,1) == UnqDate(ii);
    eval(['dataOut.d' num2str(ii) ' = FxStIn(indx, 2:6);']);

      %Have tried but died
      %dataOut(ii) = FxStIn(indx, 2:6);    
      %dataOut.d([ii]) = FxStIn(indx, 2:6);
      %dataOut.([ii]) = FxStIn(indx, 2:6);

end

FxStOut = dataOut;
FxStIndex = UnqDate;

Is 'eval' the only way to achieve this, both reading in and out of array. A number/numeral can not be a variable/array name, as I would like;

data.1
data.2
data.3

is this correct.

Can I solve this by using a 3D (:,:,:), however, each depth is a different size (factors different) and I presume that is not allowed.

[Scragmore]

Additional information on how I want to expand this.

for ii = 1:length(x)

opMax = zeros(1:length(x), 1:5);

eval(['opMax(' num2str(ii) ') = max(x.d' num2str(ii) ');']);

end

Each 'x.d1' is a (1:n,1:5) array. I am trying to find the max in each column of 'x.d1' and return the five ans to the corresponding row of new array. This 'eval' and 'num2str' seems really messy and complicated, as I build up complexity I feel this will become less and less readable as code. The above test example does not work, incorrect array size allocation.

[Stephen23]

All of the answers say "avoid eval". Read this to know why eval is a really bad way of programming:

https://www.mathworks.com/matlabcentral/answers/304528-tutorial-why-variables-should-not-be-named-dynamically-eval

Answer 1

[Teja Muppirala]

You are right in that using EVAL to dynamically generate variable names is generally a messy approach that's not very readable, slow, and doesn't scale well.

One alternative is using cell arrays.

http://www.mathworks.com/help/techdoc/matlab_prog/br04bw6-98.html

Cell arrays let you store differently sized data, for example:

M{1} = [1 2 3]
M{2} = [1 2; 3 4]
M{3} = 'hello'

M{1} + 10
M{2} * 5
disp(M{3})

[David Holdaway]

Cell arrays appear a great way of doing this! Thanks very much!

Answer 2

[bym]

eval can be evil! avoid if possible.

read this for alternatives: http://matlab.wikia.com/wiki/FAQ#How_can_I_create_variables_A1.2C_A2.2C....2CA10_in_a_loop.3F

Answer 3

[Walter Roberson]

Replace your line

eval(['dataOut.d' num2str(ii) ' = FxStIn(indx, 2:6);']);

with

dataOut.(['d' num2str(ii)]) = FxStIn(indx, 2:6);

or alternately with

dataOut.(sprintf('d%d',ii)) = FxStIn(indx, 2:6);

[Scragmore]

Walter,

Thanks. Shame you cant accept two ans. Your response answered my question directly but Teja's ans gave me an alternative view and solution to my problem that I am going to run with.

Thanks All