Extending struct member with another struct

Question

Instead of making a new question I will edit this one by completely erasing the previous one, hopefully making it less confusing.

Wall of text.

I have a basic struct that has some basic values, such as image width, height and x and y positions, like so:

struct obj {
    int id = 0;
    float x = 0;
    float y = 0;
    int image_w = 0;
    int image_h = 0;
    int depth = 0;
}

I then have an initialiser function that creates the members of that struct and stores them in an array. If that array is named "instance" then individual members and their values can be accessed by simply doing this: instance[number].x etc..

Then I have a loop or two which handle all these members and do so in the order of their depth value, defined in struct and set in initialiser function. Like so (simplified):

for (i=0;i<maxdepth;i++) {
    if (instance[n].depth == i) { doStuff; }
}

In "doStuff" function I check the members' id value in a switch statement and then have them do whatever I want inside case labels; this gives me the option to have some individual behavior within the same struct. And here is where the problem is. Although this works just fine I can't have individual fixed (or starting) variables within certain members without every member having those same variables and obviously with enough members this eventually results in a struct that is simply undesirably big and has a lot of redundancy; wasted resources. E.g I want some members to have speed and direction variables but don't want to give them to static members of the same struct that don't need them.

The question is, how do I achieve this effect without changing the fundamental idea of using structs or is there a better alternative to do this?

And I'm sorry about formatting and all; this is my first question on this website.

Answer 1

My understanding of structs is that the bigger it is the more time it takes to access its individual variables.

Your understanding is fundamentally mistaken. The size of a struct has little (if any) effect on the time required to access an individual variable inside that struct.

Regardless of that, however, your basic idea of structuring the data so one struct contains (or owns a pointer to) some other structs is perfectly fine and reasonable.

What's not so fine or reasonable is making that pointer essentially un-typed so it can refer to any other type. If you want a collection of clothes, you'd probably start with a clothing base class, and then derive various other types from that (coat, shirt, slacks, jeans, etc.). Then the person type might contain (for example) a vector of pointers to clothing, so it can contain pointers to all the other types derived from clothing.

As far as "extending scope" goes...well, I can't say much beyond the fact that I can't make much sense of what you're trying to say there.