Sandwiching "num2str" in between file-naming loop.

I'm trying to create an automated loop to extract data from .txt files that are generated from a different script. <NoNodes> is describing for a variable such that NoNodes = [16, 36, 64, 100, 144 . . . (n-2)^2, n^2]. Just squaring even integers starting at 4.

The real issue I am having is solving for <myfilename(i)>. I've put a code break at line that defines <Stress> so I can see the affects. It combines strings in a way that I do not understand and generates numbers I do not know how they've been solved.

I.e. for i = 1, this is the string output I got:

ans = 
     "CombinedStresses.Periodic.N49CombinedStresses.Periodic.N54"

Any idea's for the code below?

N = 100;
NoNodes = (4:2:N).^2
A = "CombinedStresses.Periodic.N%u"; 
B = ".Z5.Gaussian.Data.10_08_2020.Trial.1.txt";

    for i = 1:length(NoNodes);
        myfilename(i) = sprintf(A, num2str(NoNodes(i)), B);
        filename = fullfile('D:\University\Fall 2020\Class\HW3\Raw Data\Stresses', myfilename);
        Stress = readtable(myfilename(i));
    end

[Walter Roberson]

no it doesn't. The num2str gives '16' which is char(49 54) and the %u converts the char(49) to '49'. The vector is not finished outputting so it repeats the format and converts the char(54) to '54'. So you end up with like 'PQR49PQR54'

Answer 1

[Walter Roberson]

N = 100;
NoNodes = (4:2:N).^2
A = "CombinedStresses.Periodic.N"; 
B = ".Z5.Gaussian.Data.10_08_2020.Trial.1.txt";

    for i = 1:length(NoNodes);
        myfilename(i) = A + NoNodes(i) + B;
        filename = fullfile('D:\University\Fall 2020\Class\HW3\Raw Data\Stresses', myfilename(i));
        Stress = readtable(filename);
    end

The problem you had was you have a %u format element in your format, which was the content of A, but you were passing in the corresponding position the content of num2str(16) . The result of num2str(16) is '16' which is char([49 54]) . The character codes were converted through the %u format item.

[Kieran Fung]

Thanks for the help Walter :)

Going through the Mathworks examples didn't quite capture what I was after but you've exposed me too a cool new possibility when comning strings. Thanks!

One question: Why did we not have to use any float operators here when summing < A + NoNodes(i) + B> here?

[Walter Roberson]

A quite valid alternative for you would have been

N = 100;
NoNodes = (4:2:N).^2
A = "CombinedStresses.Periodic.N"; 
B = ".Z5.Gaussian.Data.10_08_2020.Trial.1.txt";

    for i = 1:length(NoNodes);
        myfilename{i} = sprintf('%s%u%s', A, NoNodes(i), B);
        filename = fullfile('D:\University\Fall 2020\Class\HW3\Raw Data\Stresses', myfilename{i});
        Stress = readtable(filename);
    end

When you use " around a list of characters, you are defining a string() object, which is distinct from a character vector. string() objects have methods defined for them that are not available for character vectors.

When you have a character vectors P and Q, the P+Q operation is defined like double(P)+double(Q) -- and has the same limitations such as one of them having to be scalar or else the two having to be the same length.

>> 'ABC'+1, char(ans)
ans =
    66    67    68
ans =
    'BCD'
    

However, for string() objects P and Q, the P+Q operation is defined to be string concatenation, like

    string(strcat(cellstr(P), cellstr(Q)))

and furthermore, if one of P or Q is string and the other is numeric, then the P+Q operation is like

    string(strcat(cellstr(string(P)), cellstr(string(Q))))
    

and string() applied to a numeric value is something like num2str() of the numeric value.

    "F" + (1:5)

would construct ["F1" "F2" "F3" "F4" "F5"] which is a string array of size 1 x 5, and indexing at (1) would be "F1" .

There is no float() operation because the operating involved is concatenation not arithmetic.