«

vector<double> v1(5,1);
vector<double> v2;

v2 = v1; // 1
v2.assign(v1.begin(),v1.end()); // 2

Are 1 and 2 the same, or are their subtle differences between them.
Which is preferable, if either? And yes, I know I could use the
construction vector<double> v2(v1), but I'm giving an example above.

Thank you c++ users.

Chris Roth wrote:
> vector<double> v1(5,1);
> vector<double> v2;
>
> v2 = v1; // 1
> v2.assign(v1.begin(),v1.end()); // 2
>
> Are 1 and 2 the same, or are their subtle differences between them.
> Which is preferable, if either? And yes, I know I could use the
> construction vector<double> v2(v1), but I'm giving an example above.
>
> Thank you c++ users.

No observable difference between them, but 1 is clearly better since it
is clearer to any reader of the code. With 2 you have to check the
arguments to understand the meaning.

john

Chris Roth wrote:
> vector<double> v1(5,1);
> vector<double> v2;
>
> v2 = v1; // 1
> v2.assign(v1.begin(),v1.end()); // 2
>
> Are 1 and 2 the same, or are their subtle differences between them.
> Which is preferable, if either? And yes, I know I could use the
> construction vector<double> v2(v1), but I'm giving an example above.

As John said, there's probably no detectable difference. The latter
construct (assign) is more useful when you want to copy a subvector.

e.g.:

vector<double> v1;
vector<double> v2;

// fill v1 here.

vector<double>::iterator start_iter = some_iterator_into_v1;
vector<double>::iterator end_iter = some_other_iterator_into_v1;

v2.assign(start_iter, end_iter);

On Feb 21, 4:40 pm, red floyd <no.s...@here.dude> wrote:
> Chris Roth wrote:
> > vector<double> v1(5,1);
> > vector<double> v2;
>
> > v2 = v1; // 1
> > v2.assign(v1.begin(),v1.end()); // 2
>
> > Are 1 and 2 the same, or are their subtle differences between them.
> > Which is preferable, if either? And yes, I know I could use the
> > construction vector<double> v2(v1), but I'm giving an example above.
>
> As John said, there's probably no detectable difference. The latter
> construct (assign) is more useful when you want to copy a subvector.
>
> e.g.:
>
> vector<double> v1;
> vector<double> v2;
>
> // fill v1 here.
>
> vector<double>::iterator start_iter = some_iterator_into_v1;
> vector<double>::iterator end_iter = some_other_iterator_into_v1;
>
> v2.assign(start_iter, end_iter);

Also, using assign allows you to assign across container types:

vector<int> v;
list<int> ll;

// fill ll here

v.assign(ll.begin(), ll.end());

red floyd wrote:
> Chris Roth wrote:
>> vector<double> v1(5,1);
>> vector<double> v2;
>>
>> v2 = v1; // 1
>> v2.assign(v1.begin(),v1.end()); // 2
>>
>> Are 1 and 2 the same, or are their subtle differences between them.
>> Which is preferable, if either? And yes, I know I could use the
>> construction vector<double> v2(v1), but I'm giving an example above.
>
> As John said, there's probably no detectable difference. The latter
> construct (assign) is more useful when you want to copy a subvector.
>
In case 1, v1 must be a vector.
In case 2, v1 can be anything provided that the iterators returned
are of a type that's insertable into v2.