Plot loop n variables

Hello Community,

I have been trying to create a plot of several arrays in loops, but I have problems assigning the change of name of the variable to the plot. Also if you know how to assign the legend name according to the plot it would be great. So, I have A1,A2,A3,... and B1,B2,B3,... respectively each pair makes a plot. So far I have written: Arrays A, B are a size (5x10000), and I'm trying to plot the second column of A (date) with the second column of B with:

%nrows is the number of pairs of A and B

for i=1:nrows
    x=datetime(A{i}(:,2)', 'ConvertFrom', 'datenum');
    y=B{i}(:,2)';
    figure
    plot(x,y);
    xlabel('A');ylabel('B'); grid on; grid minor; legend ('B',num2str(i));
    hold on
end

The error that I get is "Undefine variable A or Class A"

Any help is very appreciated. Thanks

[John D'Errico]

Yet another reason why making numbered variables like this is foolish, terribly bad programming style.

Learn how to use arrays.

[per isakson]

@Jorge

• A and B are they variables alongside A1,A2,A3,... and B1,B2,B3,.. in the base workspace?
• y=B{i}(:,2) hints that B is a cell array
• this for-loop is it part of a script or a function ?
• if in a script, please run whos A B and show us the result!
• if in a function, do you pass A and B to this function?

[Stephen23]

"...but I have problems assigning the change of name of the variable to the plot.... So, I have A1,A2,A3,... and B1,B2,B3,..."

Of course you will have problems if you have numbered your variables: using numbered variables means that you are writing slow, inefficient, obfuscated, buggy, hard-to-debug code. Some other beginners will tell you that that is okay, but it is really very simple to write neat and efficient code: use indexing instead of numbered variables. All experts and experienced MATLAB users will tell you to use indexing, and you should seriously consider following their advice.

This is what the MATLAB documentation has to say about trying to access numbered variables: "A frequent use of the eval function is to create sets of variables such as A1, A2, ..., An, but this approach does not use the array processing power of MATLAB and is not recommended. The preferred method is to store related data in a single array"

source: https://www.mathworks.com/help/matlab/matlab_prog/string-evaluation.html

You should read these:

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

https://www.mathworks.com/matlabcentral/answers/57445-faq-how-can-i-create-variables-a1-a2-a10-in-a-loop

etc, etc.

In your last question you should have used the answer using accumarray, which would split your array into a cell array, and then you can access the cell array data using indexing.

DO NOT USE NUMBERED VARIABLES.

[Image Analyst]

I'm wondering if per is right and he just has two cell arrays A and B, and when he wrote A1, A2, etc. he really meant A(1}, A{2}, etc. In fact his code tries to reference A like that, but it appears the problem is that A was never defined/created/passed-in in the first place.

[Jorge Rodriguez]

Thanks for the suggestions. As you can figure out from my code. I'm new in MATLAB or coding in general, that the reason I ask.

Answer 1

[Sebastian Castro]

This is happening because the syntax A{1} doesn't mean you're accessing a variable named A1. It's accessing the first element of a variable named A.

To put John D'Errico's suggestions in a ... more positive way, the 2 syntaxes above show the difference in stringing together things in a single array vs. having to manage a bunch of hard-coded variable names like A1, A2, etc. The former approach ends up being better to work with because it's less manual effort.

Based on your previous question, if you used accumarray to split your original data you would get a cell array. This is a way to store data of non-uniform size and/or types, which was the way to handle splitting your particular dataset. If you used that approach, you can then access each individual array as A{1}, A{2}, etc.

- Sebastian

[Jorge Rodriguez]

Thanks Sebastian, I will try to implement this format.

Answer 2

[John BG]

Hi Mr Rodriguez

this is John BG <mailto:[e-mail] [e-mail]>

while it's a good general guideline to try to work with variables that use matrix indices instead of putting the index in the name of the variable, it's a common practice, it happens, and since you already have the data in such format, let's work with what we've go available:

1.

Let's say you have the following

A1 =
    1.4897    1.4172
    1.4090    0.6715
>> A2
A2 =
   -1.2075    1.6302
    0.7172    0.4889
>> A3
A3 =
    1.0347   -0.3034
    0.7269    0.2939
>> B1
B1 =
   -0.8095    1.4384
   -2.9443    0.3252
>> B2
B2 =
   -0.7549   -1.7115
    1.3703   -0.1022
>> B3
B3 =
   -0.2414    0.3129
    0.3192   -0.8649

2.

now capture the names of the variables of interest

S=whos
S = 
  10×1 struct array with fields:
    name
    size
    bytes
    class
    global
    sparse
    complex
    nesting
    persistent

S.name
ans =
    'A'
ans =
    'A1'
ans =
    'A2'
ans =
    'A3'
ans =
    'AB'
ans =
    'B1'
ans =
    'B2'
ans =
    'B3'
ans =
    'S'
ans =
    'ans'

3.

Now have the choice to either

3.1.- work with the names of the variables contained in, for instance

D={S.name}

D =
  1×10 cell array
    'A1'    'A2'    'A3'    'AB'    'B1'    'B2'    'B3'    'D'    'S'    'ans'

there may not be need for a cell, but if you have more than 9 pairs A B then the length of the variable names varies with the increasing numerals

3.2.- or use the code you have already included in your question

A={}
for k=1:1:length(D)
if findstr(D{k},'A')
    A=[A;D{k}]
end
end

B={}
for k=1:1:length(D)
if findstr(D{k},'B')
    B=[B;D{k}]
end
end

now you can start working on the plots mentioned in the question, or any other processing

if you find this answer useful would you please be so kind to consider marking my answer as Accepted Answer?

To any other reader, if you find this answer useful please consider clicking on the thumbs-up vote link

thanks in advance

John BG

<mailto:[e-mail] [e-mail]>

[Stephen23]

Note that using whos (or any other introspective programming) is highly inefficient and is not recommended by any MATLAB experts or by the MATLAB documentation: "Avoid functions that query the state of MATLAB such as inputname, which, whos, exist(var), and dbstack. Run-time introspection is computationally expensive"

source: https://www.mathworks.com/help/matlab/matlab_prog/techniques-for-improving-performance.html

Using this method will also lead very inefficient accessing of those variables, which is not recommended by any MATLAB experts or by the MATLAB documentation: "A frequent use of the eval function is to create sets of variables such as A1, A2, ..., An, but this approach does not use the array processing power of MATLAB and is not recommended"

source: https://www.mathworks.com/help/matlab/matlab_prog/string-evaluation.html

Read this to know more about why this is a buggy, inefficient way to write code, and what the better alternatives are (hint: indexing):

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

[Jan]

Stephen hits the point again. The problem of numbered names has been discussed very often already and avoiding them is strongly recommended.

The iterative growing of arrays in the paragraph "3.2" in the above answer is very inefficient and a standard case for a bad programming pattern.

[Walter Roberson]

If you whos a symbolic variable then MATLAB will ask the symbolic engine what the size of the variable is, which can take a long time to compute in some cases. I hit a situation where it took whos about 7 minutes to figure out the size of a symbolic variable. The size information is not cached on the MATLAB side, so that was 7 minutes every time that the size was asked for.

[John BG]

it will be 'long' as in 'langsam' if the name of the variable is complicated and there are many complicated variable names.

But how complicated and slow can it be to retrieve A1 A2 .. and B1 B2 ..

People come on, when Coberdick and Simon say 'slow' without bringing any time measurement, what's the point?

[Jan]

"Slow" means, that using arrays instead of a set of variables with numbered names will be "faster" - measured in runtime and taking into account the the time for programming, debugging and maintaining the code.

You find enough arguments to get the point e.g. in https://www.mathworks.com/matlabcentral/answers/304528-tutorial-why-variables-should-not-be-named-dynamically-eval.