Assign Value to Multiply Indexed Range

Suppose control is an m-element one-dimensional structure array such that each element contains another n-element one-dimensional structure array W, with a scalar (specifically, LOGICAL) field status. As a simple example consider the synthetic construction

control = repmat(struct('W', struct('status', { false, true, true, false, true })), [1, 5])

In other words, the expression

control(i).W(j).status

is valid for all i=1:5 and all j=1:4.

My question is if there is an easy way of assigning W(3).status for all i=3:5 (i.e., a specific j value for a subset of the i range) without using the obvious loop

for i = 3:5, control(i).W(3).status = false; end

I didn't see a good way of persuading deal to do this.

Note: I'm perfectly willing to change my data structure if that's what it takes to simplify the task. Maybe turning W into a matrix is the right choice here (e.g., [control.W(3:5,2).status] = deal(false)). In my general case, though, the number of W elements might be different for each value of i.

Answer 1

[Bruno Luong]

You might store in a table, which by design put the two variables/row dimensions to the same level; so you can index them either way.

The struct as you built is hierachical. The advantage is you can put scatted length data in the nested j-level. But it prevents user to assign the same nest index for subset of outer index.

c = repmat({[false;true;true;false;true]},1,5)
ControlStatus = table(c{:})

Assign

i = 3:5;
j = 3;
ControlStatus(j,i) = repmat({false},size(i))

[Bård Skaflestad]

> You might store in a table, which by design put the two variables/row dimensions to the same level; so you can index them either way.

Hm, that's an interesting thought. I haven't really used tables much, but this might be a good opportunity to learn and experiment with them.

> The struct as you built is hierachical. The advantage is you can put scatted length data in the nested j-level.

Truth be told, that was the original reason for using the hierarchical structure.

> But it prevents user to assign the same nest index for subset of outer index.

Absolutely, and that's the problem I'm running into now. I think it's a good time to rethink the greater task from the ground up. Thanks for the hint!