MATLAB Answers

Merging two categorical columns using a rule

4 views (last 30 days)
Hi,
I am wanting to merge/add 2 categorical columns but need to use a rule.
For example a= [Flowering not not not], b=[not not not budburst] and I want the final output to be c=[Flowering not not Budburst]?
Basically, when merging/adding the two vectors, I want the word "flowering" or "budburst" to take precedence over "not". "Flowering" will never clash with "budburst" and "not "not" would give a "not".
Thanks, Wendy

  0 Comments

Sign in to comment.

Accepted Answer

Ameer Hamza
Ameer Hamza on 11 May 2018
Edited: Ameer Hamza on 11 May 2018
Here is one way
% Extracting data from yor file
data = readtable('phenology.xls');
a = categorical(data.A);
b = categorical(data.B);
% preparing default output vector
c = categorical(repmat({'not'}, size(a)));
indA = ~ismember(a, 'not');
indB = ~ismember(b, 'not');
c(indA) = a(indA);
c(indB) = b(indB);
isequal(c, data.C)
ans =
1 <--- The generated c is same as in your xls file.

  0 Comments

Sign in to comment.

More Answers (3)

Greg
Greg on 11 May 2018
Make your categorical array ordinal (with "not" as the lowest value), then use c = max(a,b). This is only guaranteed to work since you said valid values won't conflict.

  3 Comments

Wendy Cameron
Wendy Cameron on 11 May 2018
This sounds very elegant and I think would work but I'll now have to work out how to make a categorical array ordinal. Ultimately I will have categorical columns as part of a table with dates and temperatures.
Greg
Greg on 12 May 2018
Have a read of the documentation page " Ordinal Categorical Arrays ."
Basically, you specify the valueset (allowable entries) in order of smallest to largest:
valueset = {'not';'Budburst';'Flowering'};
var = categorical(a,valueset,'ordinal');
Wendy Cameron
Wendy Cameron on 12 May 2018
Thank you, I've read that now and it seems to make even more sense the way you've written it here!.
Thank you very much. Wendy

Sign in to comment.


Wendy Cameron
Wendy Cameron on 11 May 2018
Hi,
I have attached some sample data. Basically I have columns A and B but want to get column C. I couldn't quite get the method above to work but I am very new to Matlab.
Thanks for your patience. Wendy

  1 Comment

Ameer Hamza
Ameer Hamza on 11 May 2018
I just tested my code with your data and it is giving correct output. Can you tell which line is creating the error? Refer to my edited answer to see how to read data from the xls file.

Sign in to comment.


Wendy Cameron
Wendy Cameron on 11 May 2018
Yes that worked with the table I sent you as you describe. I get a 42 x 1 categorical - thank you very much. I'm now just trying to join that column into my original table. Join or interjoin might work.
Thanks, Wendy

  3 Comments

Ameer Hamza
Ameer Hamza on 11 May 2018
You can add the new column to the original table using,
data.newColumnName = c;
where data is the table, newColumnName is the name you want to give to the new column and c is the result.
Wendy Cameron
Wendy Cameron on 11 May 2018
Brilliant, thank you - that works perfectly. I've learnt a lot thank you.
Very much appreciated.

Sign in to comment.

Sign in to answer this question.

Tags