Passing vector of vectors to function

I have a function which accepts a

vector<vector<MyClass>> 

and modifies the MyClass instances. It's been a long time since I've written any C++ and I'm having difficulty remembering what is sufficient here for passing the entire arg by reference instead of by value.

My original method signature was:

void modifyVectorOfVectors(vector<vector<MyClass> > vec) { ... }

I want to make this memory efficient so I originally changed this to:

void modifyVectorOfVectors(vector<vector<MyClass*> > vec) { ... }

Then I realized that this would mean that my vec value would still make copies of all the inner vectors. So I changed my function signature to:

void modifyVectorOfVectors(vector<vector<MyClass*> >* vec) { ... }

Is this sufficient, or do I also need to do something like:

void modifyVectorOfVectors(vector<vector<MyClass*>* >* vec) { ... }

Could someone highlight the memory differences between all of these? Thanks a lot!

2 Answers

Simply

void modifyVectorOfVectors( vector< vector< MyClass > >& vec) { ... }

1. void modifyVectorOfVectors(vector<vector<MyClass> > vec) { ... }Copy external vector which has all elements of inner vector
2. void modifyVectorOfVectors(vector<vector<MyClass*> > vec) { ... }Copy external vector which has all inner vectors which holds all pointers of MyClass
3. void modifyVectorOfVectors(vector<vector<MyClass*>* > vec) { ... }Copy external vector which has all pointers of inner vectors and and inner vectors holds pointers of MyClass
4. void modifyVectorOfVectors(vector<vector<MyClass*> *> vec*) { ... }Copy pointer of external vector which has all pointers of inner vectors and and inner vectors holds pointers of MyClass