(English is not my native language; please excuse typing and grammar errors.)

I'm trying to create a vector<int> object with known length n.

I knew that I could do this by vector<int> v(n); or vector<int> v = vector<int>(n);. However, when I tried to do it by vector<int> v = n;, I got an Compile Error.

In my previous experience, vector<int> v = n seems the same as vector<int> v = vector<int>(n), but it proves that I'm wrong.

I've read the cpp reference and searched "C vector initialize with an integer" on stackoverflow but cannot find much useful information.

So what's the difference between the three ways? Thanks in advance.

CodePudding user response：

vector<int> v(n) generate a vector named "v"

vector<int> v declare a vector named "v"

vector<int> v = vector<int>(n) means that you generate a temp vector<int>(n) and v = temp . "v" and "temp" have the same type vector<int> so you can use "=" on them.

But vector<int> v = n is vector<int> = int they don't have the same type.

CodePudding user response：

Case 1

Here we consider the statement:

vector<int> v = n;  //this is copy initialization

The above is copy-initialization. But the constructor for std::vector that take size as argument is explicit and hence cannot be used here, and so this fails with the error that you're getting.

Case 2

Here we consider the statement:

vector<int> v = vector<int>(n); //this is also copy initialization

The above is also copy initialization. But this time, there is a copy constructor of std::vector that takes a vector as an argument and so this works without any error. Here the vector named v is created as a copy of(prior C 17) the temporary vector on the right hand side.

Also note that from C 17 onwards, due to mandatory copy elison, it is guaranteed that v is constructed directly using the ctor that takes an size as argument instead of being created as a copy using the copy ctor.

Case 3

Here we consider the statement:

vector<int> v(n); //this is direct initilaization

The above is direct initialization and it creates a vector named v of size n. This works because even though the ctor that takes size as argument is explicit, it can be used in direct initialization.