For this particular application, I wouldn't use the dot notation to access the variables in your table. I'd use one of the other ways to access data in a table, specifically curly brace indexing, to extract the contents of the variable.
load patients
T = table(LastName, Age, Weight);
head(T) % Show just the first few rows
If I just wanted to get all the patients whose Age is 38, I could do so in a couple different ways. I could hard-code that I want to use the Age variable:
A1 = T.Age == 38;
or I could use a more general indexing approach.
variableName = 'Age';
A2 = T{:, variableName} == 38;
isequal(A1, A2)
So if I wanted to select those patients who are 38 and weigh 131 pounds, I could hard code those constraints.
selectedPatients = T.Age == 38 & T.Weight == 131;
Or I could use that curly brace indexing. Let's start off with our criteria and a vector indicating all patients are still eligible for selection (as an initial condition.)
criteria = {'Age', 38, 'Weight', 131};
eligible = true(height(T), 1);
Now iterate.
for whichCriterion = 1:2:length(criteria)
name = criteria{whichCriterion};
value = criteria{whichCriterion+1};
eligible = eligible & T{:, name} == value;
end
Compare the two approaches to check:
isequal(selectedPatients, eligible)
This approach doesn't care how long your criteria cell array is (so long as it has an even length.) It gets a little more complicated if you wanted to use different comparisons (less than, greater than, etc.) instead of or in addition to equality testing, but not significantly more so.
selectedPatients = T.Age > 35 & T.Weight < 150;
Here the criteria is a cell array with three elements: name, comparison method, and value to compare the variable value against.
criteria = {'Age', @(x, y) x > y, 35, 'Weight', @(x, y) x < y, 150};
eligible = true(height(T), 1);
Now iterate.
for whichCriterion = 1:3:length(criteria)
name = criteria{whichCriterion};
comparisonMethod = criteria{whichCriterion+1};
value = criteria{whichCriterion+2};
eligible = eligible & comparisonMethod(T{:, name}, value);
end
Compare the two approaches to check:
isequal(selectedPatients, eligible)
Instead of using the anonymous functions I could have just used the names of the functions associated with the relational operators. gt, > is greater-than and lt, < is less-than.
criteria = {'Age', @gt, 35, 'Weight', @lt, 150}
eligible = true(height(T), 1);
for whichCriterion = 1:3:length(criteria)
name = criteria{whichCriterion};
comparisonMethod = criteria{whichCriterion+1};
value = criteria{whichCriterion+2};
eligible = eligible & comparisonMethod(T{:, name}, value);
end
isequal(selectedPatients, eligible)
Note that while I created criteria manually, you could do it as inputs to a function. Compare what the following function shows it receives as input with the criteria variable above.
showInputsAsCell('Age', @gt, 35, 'Weight', @lt, 150)
function showInputsAsCell(varargin)
disp(varargin)
end
