Select last numeric character(s) of the string

Hello,

I'm hoping for an advise on the code which has to do with extracting the digits (either 1,2 or 3) from the string which includes different condition names. It's about a rating from -100 to 100 given to different conditions.

In our other script the output gave the ratings saves as: e.g "sucrose: 10". Now it's: e.g. "sucrose10".

How can I extract the last numbers occurring at the end of different strings? I'll provide the code we had below to deal with the separation using ":" which now doesn't work.

log.rating = zeros(height(log),1);

for l = 1:height(log)
    
    if contains(char(log.Code(l)),'score','IgnoreCase',true)
        scorestring = char(log.Code(l));
        
        if strcmp(scorestring(end-3:end),'-100')
            log.rating(l) = str2double(str(scorestring(end-3:end)));
            
        elseif ~contains(scorestring(end-2:end))
            log.rating(l) = str2double(str(scorestring(end-2:end)));
            
        else
            log.rating(l) = str2double(str(scorestring(end-1:end)));
            
        end
        
    else
        log.rating(l) = NaN;
        
    end
    
end

Many thanks!!

[Aleksandra Budzinska]

The example strings are: 'Sucrose low10','Sucrose high-100','Erythritol high5','Erythritol low-10', 'Control0'

Answer 1

[Stephen23]

"How can I extract the last numbers occurring at the end of different strings?"

X = "sucrose-10";
Y = str2double(regexp(X,'[-+]?\d{1,3}$','match','once'))

Y = -10

This replaces your innermost IF-ELSEIF-ELSE-END. You could probably use it to replace you entire FOR loop too.

It detects from 1 to 3 digits at the end of the text.

It will also detect an optional +/- sign in front of those digits.

See also:

https://www.mathworks.com/help/matlab/matlab_prog/regular-expressions.html

https://www.mathworks.com/matlabcentral/fileexchange/48930-interactive-regular-expression-tool

[Stephen23]

Assuming that LOG is a table (you did not explain this in your question):

log = array2table(["scoreABC-23";"DEF 100";"scoreGHI-100";"JKLscore: 99"], 'VariableNames',{'Code'})

log = 4×1 table
         Code     
    ______________

    "scoreABC-23" 
    "DEF 100"     
    "scoreGHI-100"
    "JKLscore: 99"

idx = contains(log.Code,'score','IgnoreCase',true);
log.Rating = nan(size(log.Code));
log.Rating(idx) = str2double(regexp(log.Code(idx),'[-+]?\d{1,3}$','match','once'))

log = 4×2 table
         Code         Rating
    ______________    ______

    "scoreABC-23"       -23 
    "DEF 100"           NaN 
    "scoreGHI-100"     -100 
    "JKLscore: 99"       99 

Don't fight MATLAB with loops and IFs. Think in terms of vectors and matrices.

[Aleksandra Budzinska]

Dear @Stephen23 the first sugegstion worked very well and indeed I got rid of the IF-ELSEIF-ELSE-END parts. Thank you very much!!

Answer 2

[Kevin Holly]

output = 'sucrose: 10';

numbers = str2double(extract(output, digitsPattern))

numbers = 10

output = "sucrose10";
numbers = str2double(extract(output, digitsPattern))

numbers = 10

[Aleksandra Budzinska]

Dear Kevin,

Thanks for your quick reaction although I don't understand the difference between your first and second code.

I was trying to use for all the possible lengths of the digits we might have but it doesn't work properly:

if strcmp(scorestring(end-3:end),'-100')
    log.rating(l) = str2double(extract(scorestring(end-3:end)));
    
elseif ~contains(scorestring(end-2:end))
    log.rating(l) = str2double(extract(scorestring(end-2:end)));
    
else
    log.rating(l) = str2double(extract(scorestring(end-1:end)));

[Dyuman Joshi]

"although I don't understand the difference between your first and second code. "

@Aleksandra Budzinska, they are the same code, Kevin just used two different inputs to show the result.

However, @Kevin Holly, this fails to include the negative sign for negative numbers -

output = 'Sucrose high-100';

numbers = str2double(extract(output, digitsPattern))

numbers = 100