Define a name of the variable using the value of other variable

Suppise I have data

Month Day Visitors
----------------------------
January 1 100
January 7 50
February 1 400
February 3 200

I want to create a variable that is a proxy for months such as

Month Day Visitors proxy_January proxy_February
------------------------------------------------
January 1 100      1             0
January 7 50       1             0
February 1 400     0             1
February 3 200     0             1

One way to do this is to define variables one by one for each month.

But I would like to know how I can let Matlab do this automatically.

I start with

set_of_month=unique(T.Month)

that will give me the set of months. Then I can iterate over the set of months.

But how can I define a variable using set_of_months? I like to do something like

for i=1:length(set_of_states)
    for j=1:height(T)
        if T.Month(j)=set_of_month(i)
            T.set_of_month(i)=1
        else
            T.set_of_month(i)=0
        end
    end
end

Please advise.

[Walter Roberson]

Are the months categorical or cell array of character vectors?

[alpedhuez]

cell array

Answer 1

[Marco Riani]

Please let me know if I understood correctly your question.

Month={'January'; 'January'; 'February'; 'February'};
set_of_month=unique(Month);
% Add a matrix of zeros to host the dummy variables
dum=zeros(length(Month),length(set_of_month));
T=[table(Month) array2table(dum)];
% Add the labels
T.Properties.VariableNames(2:end)=set_of_month;
% Add the ones in the corresponding position
 for i=1:length(set_of_month)
    for j=1:height(T)
        if strcmp(T.Month{j},set_of_month{i})
            T{j,i+1}=1;
        end
    end
 end
 
 disp(T)
%       Month        February    January
%     ____________    ________    _______
% 
%     {'January' }       0           1   
%     {'January' }       0           1   
%     {'February'}       1           0   
%     {'February'}       1           0   

[alpedhuez]

Let me work on it.

[alpedhuez]

"You cannot assign numeric values to a datetime array."

[alpedhuez]

Got it. Thank you very much.

Answer 2

[Walter Roberson]

 T.(set_of_month{i}) = value

[alpedhuez]

"Table variable names must be strings or character vectors." Please advise.

[alpedhuez]

There are two issues.

First, how can one define a variable of a table with the name taken from set_of_months?

Second, how can one compare a cell array values for if conditional?

[Walter Roberson]

The implication is that set_of_month is not a cell array of character vectors as you had claimed.

Perhaps your month entries in the table are individual cells rather than character vectors. If so then

monthname = set_of_month{i}{1}; 
T.(monthname) = value

and monthname is something that you can strcmp() or ismember()

[alpedhuez]

Let me work on it.

Answer 3

[Steven Lord]

M = repelem(["January"; "February"], 2, 1);
D = [1; 7; 1; 3];
V = [100; 50; 400; 200];
T = table(M, D, V, 'VariableNames', ["Month", "Day", "Visitors"])

T = 4x3 table
      Month       Day    Visitors
    __________    ___    ________

    "January"      1       100   
    "January"      7        50   
    "February"     1       400   
    "February"     3       200   

UM = unique(T.Month);
for whichMonth = 1:numel(UM)
    T.("proxy_" + UM(whichMonth)) = T.Month == UM(whichMonth);
end
T

T = 4x5 table
      Month       Day    Visitors    proxy_February    proxy_January
    __________    ___    ________    ______________    _____________

    "January"      1       100           false             true     
    "January"      7        50           false             true     
    "February"     1       400           true              false    
    "February"     3       200           true              false    

Alternately the unstack function may do what you want (or something similar to what you want.)