HELP REQUIRED IN PROGRAMMING

if i want to give; d1,d2,d3......d90 all having value as 2 i.e; d1,d2,d3....d90=2 P1,P2,P3......P50 all having value as 18 i.e. P1,P2,P3....P50=18 then how to give these statement in MATLAB

[Jan]

Creating variables with indices hidden in the name is a bad programming patter. See:

• FAQ: Create variables A1, A2, ...
• Tutorial: Do not EVAL
• MATLAB documentation: Why Avoid the eval Function?

A method to create "numbered variables" is less useful than using a clean programming style by using arrays and indices.

@adarsh: Your other questions contains code, which suffers from the same problem: Question 378029, Question 383271, Question 378029, Question 376013. There we find the same programming pattern of numbered variables. It is time to learn a better method to write clean and efficient code. Use arrays instead of a complicated bunch of numbered variables.

Answer 1

[John D'Errico]

Sorry. But this is just a bad idea. A novice style or programming, only headed for worse things to come.

What will you do with those variables? Probably want to create new numbered variables, all on the fly. So you will find ways to create that variables too. UGH. Bad programming. Bad news. Inefficient, buggy code. I'm sorry to have to tell you this, but you will be sorry in the end.

Instead, learn to use arrays and vectors. Learn to use MATLAB. What you are doing is thinking in terms of a spreadsheet. MATLAB is not a spreadsheet. Spreadsheets make it simple to do what you want, because that is how they are designed to work. But instead, there are far easier ways.

d = repmat(2,1,90);

DONE. One line. If you want to access the 17th element of that vector, simple. d(17). If you want to access multiple elements of d, then use something like d(1:2:end) for example, which will access every other element.

Or create d as

d = 2*ones(1,90);

Create P in one of those same ways.

Your code will improve by leaps and bounds, once you start to use MATLAB as it is designed to be used, instead of trying to use MATLAB as a spreadsheet.

[adarsh]

ok, that means, i need to give as:

 d=repmat(2,1:90)
 P=repmat(18,1:50);

[Steven Lord]

Close but not quite. Those commands would fail because they would try to create VERY large arrays (the first would have more elements than there are atoms in the universe!) Take a look at the examples in the documentation for repmat and use them as a model. You can open the documentation for repmat using the command:

doc repmat

[Walter Roberson]

Not colon in the repmat, comma instead, like John showed.

[John D'Errico]

Yes. Be vewwwwy careful in how you use repmat here. Unless you have a LOT of memory. ;-) That comma I used was important. I can see what you were thinking when you write it with a colon, but it won't work that way.

Really, what you need to do is get used to writing MATLAB, using vectors and arrays. The alternative is as I said, to continue thinking about MATLAB as a spreadsheet. It is not that. Spreadsheets have great tools to deal with essentially numbered variables. But MATLAB has even better tools for using vectors and arrays. And that is what makes MATLAB truly great. It is when you try to think too deeply in terms of spreadsheets, living in the spreadsheet universe, while still trying to use MATLAB that things get hairy.

So use vectors and arrays. Your code will be hugely better for doing so.

[Steven Lord]

The array size limit preference can help protect you against the medium case where your array is big but not impossibly big. The fact that MATLAB won't even try to allocate an array with more elements than the number of atoms in the universe will save you in the impossibly big case.

[Walter Roberson]

At least twice a year, a user gets upset that MATLAB will not allocate arrays with more elements than the number of atoms in the universe. When they ask for zeros(1:3000) then, By Golly! they expect to be obeyed!

Answer 2

[John BG]

Hi Adarsh

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

there's a very easy way to obtain the variables you need with command evalin:

1. Define the following parameters

What variable strings to use:

str1={'d';'P';'sbt'}

What range for each variable

range=[9;5;12]

What value for each set of variables

val=[2;18;21]

the above values are just example, change accordingly

2.

Because the range of one or more variables can easily exceed 10, the simple approach

str_rng=num2str([1:1:range(k)]'); % do not use

cannot be used because it adds a blank ahead of the numeral for the values [1:9]

3.

Using a cell to build strings for evalin

for k=1:1:numel(str1)
    str_rng={};
    for s=1:1:range(k)
        str_rng=[str_rng;num2str(s)];
    end

      for s=1:1:range(k)
          evalin('base',[str1{k} str_rng{s} '=' num2str(val(k))])
      end

end

Now in your workspace you have all the variables generated as required.

Comment:

Some experts discourage the use of evalin in favor of building matrices containing all relevant data. That is particularly useful when reading some one else's code.

Yet I agree with you that sometimes there's need to generate independent variables as you have described in your question.

Adarsh

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 for time and attention

John BG

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

[Jan]

Not only "some" experts, but all experts and the documentation of Matlab recommend not to create variables dynamically. See the exhaustive explanation Tutorial: Do not EVAL and MATLAB docs: Why Avoid the eval Function?

range(k) is meaningless here, because k is a scalar inside the loop. The editor shows a MLint-warning because of the iterative growing of str_rng - serious code would use a pre-allocation. evalin('base') creates the variables in the base workspace, where they are not needed. Therefore this is a bad implementation of a bad idea.

[Stephen23]

Compare this suggestion from John D'Errico:

d = repmat(2,1,90);

with John BG's proposal:

for k=1:1:numel(str1)
    str_rng={};
    for s=1:1:range(k)
        str_rng=[str_rng;num2str(s)];
    end
      for s=1:1:range(k)
          evalin('base',[str1{k} str_rng{s} '=' num2str(val(k))])
      end
end

Now some simple questions:

• Which of these will be slower?
• Which of these will be harder to debug?
• Which of these will waste more of the programmer's time?
• Which of these makes further processing of that data more difficult?
• Which of these requires multiple functions, temporary variables, and for loops, just to store exactly the same data?
• Which of these totally destroys any possible JIT code acceleration, thus forcing beginners into writing much slower code?

[Jan]

@adarsh and other readers: This is a bad choice of an accepted answer. John BG's code does not create the wanted variables where they are needed and even if you modify it to work properly, the eval methods are known to create more problems than they seem to solve. This topic is discussed several times each week and everybody but John BG suggests better, cleaner and more efficient solutions.

Several experienced programmers offered you help which is useful in a long term view. And John BG's eval suggestion means drilling a hole in your knee. Of course, it is your choice.

[Guillaume]

Chalk another one for the "this is not a good answer" club. There may be some rare cases where eval may be useful. The present case is absolutely, definitively, not one of them. Creating 90 d and 50 P variables is complete madness.

[John D'Errico]

It is madness, because the next step is what will you do with all those crap variables? The answer is simple. They will need to create OTHER variables, with similar names. So they will now be creating E1-E90, and Q1-Q50 etc. Or they will want to create sums of those variables. So they will be asking how to perform a sum over all those variables. Etc. Quick steps into pure programming hell.

We can say thanks to John BG for helping to bring yet another person into a programming style that will only earn the ire of anyone who ever tries to use or even read their code from hell. And of course, when adarsh tries to debug their own code, expect many more frenzied questions, about how do I debug this crappy code. Given that the alternative is so trivially simple to write and use ... way to go John BG.