Currently, my code is something like this:
vector<vector<string>> result;
for (Example e : someExampleIterable) {
result = e.evaluate(result);
}
// Do some operations with the result later
vector<vector<string>> Example::evaluate(vector<vector<string>> intermediate) {
vector<vector<string>> result;
vector<vector<string>> newTable = someExternalAPI.get(someAttributeOfThisObject);
for (vector<string> row : intermediate) {
// Some complex operation that combines and manipulates data from newTable & intermediate
result.push_back(final_data);
}
return result;
}
As you can see what I'm trying to achieve is like a "pipeline" sort of loop where I constantly update my result.
Since my function that manipulates the existing table involves creating a new table and returning it due to the complex operations done with the newTable and intermediateTable (plus this is how I coded it to have this pipeline), is there a way where I can use references so that I do not have to keep on shallow copying my vector<vector> which I understand is expensive in C++, especially since my intermediate can be quite large.
Would changing the function definition to the below work even with the reassigning? Or what is the best practice for this kind of pipelining kind of loop.
vector<vector<string>> Example::evaluate(vector<vector<string>>& intermediate)
If possible, I would want to do a:
void Example::evaluate(vector<vector<string>>& intermediate)
where I can manipulate my intermediate table directly but as mentioned, the operations are quite complex and it would be easier to do on a new empty table and return it.
Thanks!
