Assign a numerical value based on the value of a string.

I have data which contains a series of text strings. I have read this in from a CSV file.

The data can be thought of as "Good", "Better", "Best". I want to assign a numerical value to these so I can do some quantitative analysis.

I thought a good first step would be to create a vector with ['Good'; 'Better'; 'Best';] in and then a vector which is [1;2;3].

For each given isntance of "Good" in the data I could lookup the row ID in the first vector and then match to the row ID in the second vector.

But this feels clunky and in any case it seems you cannot put different length strings into each row of a vector - it returns

Error using vertcat

Dimensions of arrays being concatenated are not consistent.

[Stephen23]

The MATLAB approach: https://www.mathworks.com/help/matlab/categorical-arrays.html

Answer 1

[Karl Cronburg]

Fixed Set of Enumerated Strings

If the strings in the text file are known ahead-of-time, it may make sense to assign each string a numerical value and then compare against that value any time you need to check which string you have:

classdef Goodness
    properties (Constant = true)
        Good = 1;
        Better = 2;
        Best = 3;
    end
end

And then parse your strings into the Goodness values:

function goodness = stringToGoodness(str)
    switch str
        case 'Good'
            goodness = Goodness.Good;
        case 'Better'
            goodness = Goodness.Better;
        case 'Best'
            goodness = Goodness.Best;
    end
end

Now you can do things like this elsewhere in your code to check which kind of goodness value you are looking at:

function processData(myData)
    for idx = 1:numel(myData)
        if myData(idx) == Goodness.Good
            % My code for handling data that is good / OK.
        elseif myData(idx) == Goodness.Better
            % My code for handling data that is better than just 'good'.
        end
    end
end

Another option would be enums, although numeric literals will be more performant if your data set is large.

Key-Value Dictionary

Alternatively, instead of looking up the row ID in the first vector, you could use a dictionary like so:

dict = dictionary;
dict('Good') = 1;
dict('Better') = 2;
dict('Best') = 3;

if myData(5) == dict('Good')
    % Code for 'good' data.
end