How to filter a table in matlab and then assign a value using a loop

Hi. I have a table A with 5 columns, the first three columns are categorical and the other two are numbers. The table contains 198 riws and is like this:

TumoSize NodalStatus Grading ER HER2

pT1b pN0 G2 0 0

pT2 pN3 G3 90 3

pT2 p1mi G2 55 2

...

Now, I have to filter this table per raw and assign a score of:

• 2 when TumorSize is equal to pT2 or pT3 or pT4
• 1 when NodalStatus is equal to pN2 or N3
• 1 when Grading is equal to G3
• 1 when ER is <70
• 1 when HER2 is equal to 0 or 1 or 2

At the end, I have to sum all this score per raw and then divide in three groups, according to the score (score between 0-2; score 3 or 4 and score 5-7)

My code, that doesn't work, is:

for r = 1:size(A,1)
    for c = 1:size(A,2)
        g = A(r,c);
        score = 0;
        if g == 'pT2' || g == 'pT3' || g == 'pT4' || g == 'pN2' || g == 'pN3'
            score = 2;
        else if g == 'G3' || g >= 0 && g<70 || g == 0 || g == 1 || g == 2 
                score = 1;
            else
                score = 0
                
            end
            B(r,c) = score 
        end
    end
end
total_score = sum(B,2) %sum score
firstgroup = sum(total_score == 0 | total_score == 1 | total_score == 2)
secondgroup = sum(total_score == 3 | total_score == 4)
thirdgroup = sum(total_score == 5 | total_score == 6 | total_score == 7)

it doesn't work!!! Help me please :)

thank you in advance

[Stephan]

Please attach your table

[France]

Here the table. thanks

[Stephan]

I get other results in my answer then Andrei - please feedback...

Also please clarify if correct Nodal Status is N3 or pN3 - since you have the one in your code and the other in your question / description.

Answer 1

[Stephan]

Hi,

try:

% Define a column for Scores
A.Score(1:size(A,1),1) = 0;
% Scores due to Tumor Size
A.Score(A.TumorSize=='pT2' | A.TumorSize=='pT3' | A.TumorSize=='pT4') = A.Score(A.TumorSize=='pT2' | A.TumorSize=='pT3' | A.TumorSize=='pT4') + 2;
% Scores due to Nodal Status
A.Score(A.NodalStatus=='pN2' | A.TumorSize=='pN3') = A.Score(A.NodalStatus=='pN2' | A.TumorSize=='pN3') + 1;
% Scores due to Grading
A.Score(A.Grading=='G3') = A.Score(A.Grading=='G3') + 1;
% Scores due to ER
A.Score(A.ER<70) = A.Score(A.ER<70) + 1;
% Scores due to HER2
A.Score(A.HER2==0 | A.HER2==1 | A.HER2==2) = A.Score(A.HER2==0 | A.HER2==1 | A.HER2==2) + 1;

%Define column for Grouping
A.Group(1:size(A,1),1) = NaN;
% Assign groups due to Scores
A.Group(A.Score>=0 & A.Score<=2) = 1;
A.Group(A.Score>=3 & A.Score<=4) = 2;
A.Group(A.Score>=5 & A.Score<=7) = 3;

Best regards

Stephan

[France]

Thank you Stephan.

Best Wishes

Answer 2

[Andrei Bobrov]

Let T - your table

x = sum([ismember(T{:,1:3},categorical({'pT2','pT3','pT4','pN2','N3','G3'})),...
                                          [T.ER < 70 ,ismember(T.HER2,0:2)]],2);
T.group = discretize(x,[0 3 5 7],'categorical',...
                                     {'firstgroup','secondgroup','thirdgroup'});

[Stephan]

Nice! Does this one consider that for Turmor Size there are 2 Score points instead of 1?

[France]

Thanks a lot Andrei. Unfortunately, it doesn't consider 2 point score for Tumor Size.

[Andrei Bobrov]

Hm :), small typo. Fixed.

x = [ismember(T{:,1:3},categorical({'pT2','pT3','pT4','pN2','N3','G3'})),...
                   [T.ER < 70 ,ismember(T.HER2,0:2)]]*[2;ones(size(T,2)-1,1)];
T.group = discretize(x,[0 3 5 7],'categorical',...
                                   {'firstgroup','secondgroup','thirdgroup'});

[France]

Thank you Andrei. I get this error

Error using  * 
Incorrect dimensions for matrix multiplication. Check that the number of columns in the first matrix
matches the number of rows in the second matrix. To perform elementwise multiplication, use '.*'.

Error in Other_solution (line 8)
x = [ismember(A{:,1:3},categorical({'pT2','pT3','pT4','pN2','pN3','G3'})),...
 

[Andrei Bobrov]

Hi!

I'm use your mat-file and my code working.

>> load('J:\Octavework\answers\mat-files\A.mat')
>> T = A;
x = [ismember(T{:,1:3},categorical({'pT2','pT3','pT4','pN2','N3','G3'})),...
                   [T.ER < 70 ,ismember(T.HER2,0:2)]]*[2;ones(size(A,2)-1,1)];
T.group = discretize(x,[0 3 5 7],'categorical',...
                                   {'firstgroup','secondgroup','thirdgroup'})
T =
  198×6 table
    TumorSize     NodalStatus    Grading    ER    HER2       group   
    __________    ___________    _______    __    ____    ___________
    pT1b             pN0           G2        0     0      firstgroup 
    pT2              pN0           G3       90     3      secondgroup
    pT2              PN0           G2        0     0      secondgroup
    pT1c             pN1mi         G2       90     0      firstgroup 
    pT1b             pN0           G3       90     2      firstgroup 
    pT1c             pN0           G3       90     1      firstgroup 
    pT1b             pN0           G2       90     0      firstgroup 
    pT1a             pN0           G3        0     3      firstgroup 
    pT1c             pN0           G3        0     0      secondgroup
    pT1a             pN0           G3       95     0      firstgroup 
    

in live editor: