Renaming categories with accents

I have a categorical array t and some categories can have diacritic/accents, such as circumflexes. I want to standardize everything with no diacritic/accents.

I tried this code:

str = {'Á', 'É', 'Í', 'Ó', 'Ú','Ã','Ç','Â','Ê','Ô'};
strreplace = {'A', 'E', 'I', 'O', 'U','A','C','A','E','O'};

t = categorical({'VÉRDE','VERDE','AZUL','AMARELO','VERMELHO','VERMÊLHO'})';
cat = categories(t);
newcat = cat;

for i = 1:numel(str)
    newcat = regexprep(newcat, str{i}, strreplace{i});
end
     
B = renamecats(t,cat,newcat)

However, after removing the accents, some categories turn out to be the same, for exemple: VERMELHO AND VERMÊLHO.

So I receive the following error:

Error using categorical/renamecats (line 39)
NEWNAMES contains duplicated values.

Is there anyway around?

This is just an example. I need a very efficient code since my categorical array t is comming from a very long table with approximaly 500 categories.

Thanks,

[Jan]

500 does not sound like big data.

You did not mention what you want to happen, if the names of the categoricals are equal. So it is hard to suggest a solution.

By the way, strrep is much faster than regexprep .

[Luisa Liboni]

Jan,

My table has 27,900 lines with 12 categortical columns. One of these columns have 500 unique categories. And I need to get rid of accents in all of the 12.

VERMELHO and VERMÊLHO should be the same. So yes, after removing the accents, some categories turn out to be the same and should be merged. However, renamecats give me that error.

Answer 1

[Guillaume]

str = {'Á', 'É', 'Í', 'Ó', 'Ú','Ã','Ç','Â','Ê','Ô'};
strreplace = {'A', 'E', 'I', 'O', 'U','A','C','A','E','O'};
t = categorical({'VÉRDE','VERDE','AZUL','AMARELO','VERMELHO','VERMÊLHO'})';
cat = categories(t);

%calculation of new categories, no need for loop
newcat = replace(cat, str, strreplace);

%replace cat by newcat. Create new categorical array using newcat and the index of the original categories in t:
newt = categorical(newcat(double(t)))