Vector and matrix index operations

Hi, My question is about vectors and matrix: Assuming I have vector (array) of strings: myvect={'101AA21' '101AA22' '101AA23' '102AA21' '102AA22'}; I have a matrix mymatrix=magic(5); I want to define a variable from myvect for each column of mymatrix such the results should be: 101AA21=mymatrix(:,1); 101AA22=mymatrix(:,2), etc
in order to do this:
myvect={'101AA21' '101AA22' '101AA23' '102AA21' '102AA22'}; mymatrix=magic(5); mynew_vect=char(myvect);
for i=1:length(mynew_vect) mynew_vect(i)== mymatrix(:,i); end
plot(101AA21, 102AA22, '-r');
Statment mynew_vect(i)== mymatrix(:,i); it is wrong but how can be defined . The real vector is 300, matrix is 70000X300
Thank you

 Accepted Answer

You won't be able to vectors/matrices long with a variable name. I suggest using a table or structure for mynew_vect instead. However, the variable names you have are invalid for a structure element. They need to start with a character.
A table might have performance advantages over a structure that large. I couldn't say either way.
Using a structure:
names={'D101AA21' 'D101AA22' 'D101AA23' 'D102AA21' 'D102AA22'};
mymatrix=magic(5);
for i=1:length(names)
mynew_vect.(names{i})= mymatrix(:,i);
end
And with a table:
names={'101AA21' '101AA22' '101AA23' '102AA21' '102AA22'};
mymatrix=magic(5);
for i=1:length(names)
mynew_table.names{i}= mymatrix(:,i);
end

4 Comments

Thank you for you answer. The problem is that those names(variables) '101AA22', '101AA23' exits exact in this format. And in the final report, the customers want exact this format.
Stephen23
Stephen23 on 6 Jul 2015
Edited: Stephen23 on 6 Jul 2015
MATLAB variable names must start with a letter character: "A valid variable name starts with a letter, followed by letters, digits, or underscores"
Also you should not define variable names dynamically, but use a structure, table or cell array instead, as Keith Hooks described.
You could easily keep a cell array of those names for using in "the final report", and keep the corresponding data in a cell / table / structure. This would be much more robust, because it makes your code completely independent of how the customer names their variables. It would make your life easier.
That the customers want an illegal variable name is not relevant. They cannot have it. Variables cannot have names that start with a number.
And even if they decide to choose better names, I would STRONGLY suggest following Stephen's excellent recommendations here. It is just a bad idea to do what you wish to do.
Hi, Those variables are signals in a power plant defined in the early 1978. In order to get their licence some calculations are with specials codes. So, I cannot change the name of the variables. I write this part of the program to compare experimental data with their system code calculations. So, after 37 years of using those names for their parameters/variables they are very reluctant to any changes. Any way, thank you all. I proved all your solutions.

Sign in to comment.

More Answers (2)

Thorsten
Thorsten on 6 Jul 2015
Edited: Thorsten on 6 Jul 2015
Use eval:
myvect={'101AA21' '101AA22' '101AA23' '102AA21' '102AA22'};
mymatrix=magic(5);
for i = 1:numel(myvect)
eval(['var_' myvect{i} '= mymatrix(i,:);']);
end
plot(var_101AA21, var_101AA22)
And yes, I agree that this is awful code, but the closed to be able to write something as simple as (which seems to be what wanted here, for whatever reasons):
plot(101AA21, 101AA22)

1 Comment

DON'T use EVAL. Here's a slight modification of your code that avoids the EVAL while still getting close to the desired (illegal) code.
myvect={'101AA21' '101AA22' '101AA23' '102AA21' '102AA22'};
mymatrix = magic(5);
V = struct;
for k = 1:numel(myvect)
N = ['x' myvect{k}];
V.(N) = mymatrix(k, :);
end
plot(V.x101AA21, V.x101AA22)
Some prefix on the field name is necessary because MATLAB identifiers must start with a letter; I chose x.

Sign in to comment.

What do you think about something like this :
myvect={'101AA21' '101AA22' '101AA23' '102AA21' '102AA22'};
mymatrix = magic(5);
Tmp1 = find( strcmp( myvect, '101AA21' ) );
Tmp2 = find( strcmp( myvect, '102AA21' ) );
plot( mymatrix(:,Tmp1), mymatrix(:,Tmp2) );
But perhaps, I didn't well understand what would you want to do with this variables...

Categories

Asked:

on 6 Jul 2015

Edited:

on 25 Jun 2019

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!