Error using string concatenation as input to ordinal( ) function

Hi, I am trying to create dynamic legend titles, based on numeric input ranges (RSRP signal levels). Given input sorted_RSRP, I am adding a category column, to specify the signal ranges for plotting colour-coded map data.

To do this, I have attempted to combine string values in the ordinal( ) function so I can display the ranges (that would change for different data sets) in the legend. I get the following error:

Error using categorical (line 391)

CATEGORYNAMES must be a string array or cell array of character vectors containing non-blank category names.

Error in ordinal (line 148)

b = b@categorical(a,args{:},'Ordinal',true);

See code below:

% five data categories
p = 0:0.2:1;
breaks = quantile(rsrp,p)

% string concatenation to show numeric range values
a = num2str(breaks(1)) + " to " + num2str(breaks(2))
b = num2str(breaks(2)) + " to " + num2str(breaks(3))
c = num2str(breaks(3)) + " to " + num2str(breaks(4))
d = num2str(breaks(4)) + " to " + num2str(breaks(5))
e = "> " + num2str(breaks(5))

% categorise RSRP
catRSRP = ordinal(rsrp, {a,b,c,d,e}, [], breaks);

% I have also tried this, inputting string concatenation directly
% catRSRP = ordinal(data_sorted.RSRP, {num2str(breaks(1)) + " to " + num2str(breaks(2)), ... 
%     num2str(breaks(2)) + " to " + num2str(breaks(3)), ... 
%     num2str(breaks(3)) + " to " + num2str(breaks(4)), ... 
%     num2str(breaks(4)) + " to " + num2str(breaks(5)), ... 
%     "> " + num2str(breaks(5))}, [], breaks);

Not entirely sure what the error I am getting means, none of the a, b, c, d, e string variables are empty.

Any help would be greatly appreciated!

[the cyclist]

It would be helpful if you uploaded the data variable in a MAT file, so that we can see exactly what you are.

[Robyn Seery]

@the cyclist Apologies, I completely forgot! I have edited the question and code, adding in the data I am using.

Answer 1

[Cris LaPierre]

There is a warning at the top of the documentation page for ordinal

• The nominal and ordinal array data types are not recommended. To represent ordered and unordered discrete, nonnumeric data, use the Categorical Arrays data type instead.

I think what you want to use is discretize to assign categories to the values based on what bin they would fall into. Use the syntax

[___] = discretize(___,'categorical',categoryNames)

load rsrp.mat

% five data categories
p = 0:0.2:1;
breaks = quantile(rsrp,p);

% string concatenation to show numeric range values
a = breaks(1) + " to " + breaks(2);
b = breaks(2) + " to " + breaks(3);
c = breaks(3) + " to " + breaks(4);
d = breaks(4) + " to " + breaks(5);
e = "> " + breaks(5);

% categorise RSRP
catRSRP = discretize(rsrp,breaks,'categorical',[a,b,c,d,e]);
catT = table(rsrp,catRSRP)

catT = 1153×2 table
    rsrp      catRSRP   
    ____    ____________

    -115    -115 to -103
    -114    -115 to -103
    -113    -115 to -103
    -113    -115 to -103
    -113    -115 to -103
    -113    -115 to -103
    -112    -115 to -103
    -112    -115 to -103
    -111    -115 to -103
    -111    -115 to -103
    -111    -115 to -103
    -111    -115 to -103
    -111    -115 to -103
    -111    -115 to -103
    -111    -115 to -103
    -110    -115 to -103

[Robyn Seery]

Thank you very much, works just as I needed it to! Really appreciate your help.