Question

I have a class called Page in my program. It has 10 attributes, but at certain point, only 3 attributes are needed (RawPage) and in other point, all or almost all atributes are needed (ProcessedPage).

Is better I have only only one class with all attributes or split these class into two classes, one for every need?

EDIT: the processed atributes are new attributes in the Page, and not overwrites none.

Answer 1

I suspect someone will suggest using inheritance here, but I won’t… it sounds like you logically create a “processed page” from a “raw page” (via processing) – so why not model it exactly like that?

class ProcessedPage
{
    private final RawPage source;

    // Other stuff - the results of the processing
}

class RawPage
{
    // The data you need here

    public ProcessedPage process(...)
}

When a class has information which is only sometimes valid / needed / relevant, it ends up being quite tricky to work with, especially as the system grows in size. With a more explicit model, you do potentially end up with some duplication, but your classes end up with a better defined purpose in life, rather than suffering from multiple personality disorder.

Note that I haven’t made ProcessedPage a subclass of RawPage – because I don’t imagine you’d normally want to treat a ProcessedPage as a RawPage, potentially reprocessing it. You may want to have a common interface that they both implement which doesn’t include the process method, of course, but that’s slightly different.