Add support for applying a function to every variable within a CodimensionalCollection #3

New issue

Open

opened 2025-11-19 02:08:41 +00:00 by UserCurt · 0 comments

UserCurt commented

2025-11-19 02:08:41 +00:00

(Migrated from codeberg.org)

C = CodimensionalCollection(toeplitz(1:3), hankel(1:3), [1; 2; 3], [4, 5, 6]);

% Apply mean to each variable
R = C.apply(@mean)
R = CodimensionalCollection(mean(C.Var1), mean(C.Var2), mean(C.Var3), mean(C.Var4)); % Equivalent

% Apply mean(__, 2) to each variable
R = C.apply(@mean, 2)
R = CodimensionalCollection(mean(C.Var1, 2), mean(C.Var2, 2), mean(C.Var3, 2), mean(C.Var4, 2)); % Equivalent

The name apply is not settled. Alternatives to consider include map, transform, foreach. Though foreach doesn't suggest it returns a value so is probably not a good choice.

It may be appropriate to also apply to only a subset of variables. The apply method cannot be overloaded to account for this as it would be ambiguous whether any additional arguments should be passed in or not. Instead, a new function like applySubset could be made that that applies a mapping to a subset of variables. However, it would be unclear whether it keeps the others intact or omits them. It may be more appropriate to instead of a select method that selects a subset of variables and then subsequently using apply. This is out of scope for this feature request, but it could influence whether a method like applySubset is developed or not.

It would be quite convenient to apply a function to all variables within a CodimensionalCollection at the same time, returning a new CodimensionalCollection as a result. For example, one could perform `mean(C)` which would apply the `mean` function to each entry of the CodimensionalCollection `C`. Of course, not all methods can be known in advance, especially when `C` contains custom classes with custom methods, so it may be better to define an `apply` method: ```matlab C = CodimensionalCollection(toeplitz(1:3), hankel(1:3), [1; 2; 3], [4, 5, 6]); % Apply mean to each variable R = C.apply(@mean) R = CodimensionalCollection(mean(C.Var1), mean(C.Var2), mean(C.Var3), mean(C.Var4)); % Equivalent % Apply mean(__, 2) to each variable R = C.apply(@mean, 2) R = CodimensionalCollection(mean(C.Var1, 2), mean(C.Var2, 2), mean(C.Var3, 2), mean(C.Var4, 2)); % Equivalent ``` The name `apply` is not settled. Alternatives to consider include `map`, `transform`, `foreach`. Though `foreach` doesn't suggest it returns a value so is probably not a good choice. It may be appropriate to also apply to only a subset of variables. The `apply` method cannot be overloaded to account for this as it would be ambiguous whether any additional arguments should be passed in or not. Instead, a new function like `applySubset` could be made that that applies a mapping to a subset of variables. However, it would be unclear whether it keeps the others intact or omits them. It may be more appropriate to instead of a `select` method that selects a subset of variables and then subsequently using `apply`. This is out of scope for this feature request, but it could influence whether a method like `applySubset` is developed or not.