Data Structure for one million named objects

Binary tree

One Million Named Objects:

1. I dont think binary Tree is a good idea. Depth of the tree will be huge
2. B+ Tree is the right Data structure

Since the objects are named, I would go for the hash table.

I think the most suitable here will be a dictionary solution - the
tree where each node contains a letter. And each name will be a word
in such a dictionary, so we need the node also contains the sign that
there is a name ends here (let's say -|). For example:

N  | M | ...

I | O I | ...

K-| | L K-| | ...

| L | ...

| Y-| | ...

Since there is the question about serching and insertion quickly, i think the data structure should be binary serach tree , not simple binary tree

The ultimate solution for this is TRIE or advanced TRIE

AVL trees

B tree will be most suitable.  B+ is not because the requirement is not sequential search.

It depends on the amount of primary memory one has for the process. If we can store 1 million records in primary memory, then we can very well take a hash table with a consistent hashing scheme for avoiding collisions. It will take O(1). 

You can use a hash table or an array of objects You can use an array of objects because you already know that you need to store a specified number of objects, also you can store and access objects quickly in an array of objects. First declare the object:

Code
 
GradeBook myGradeBook = new Gradebook();
 
Then create array to contain objects (gradebooks):
 
GradeBook arrayOfObjects[] = new GradeBook[1,000,000];