Show Lecture.ContainerSizes as a slide show.

CS253 Container Sizes

Container sizes

A vector, and some other containers that use contiguous allocation, have several kinds of sizes defined:

• .size(): how many elements the vector currently contains
• .capacity(): how many elements are allocated, or, how many elements it could contain before reallocation
• .max_size(): biggest possible .size() on this computer

.size() is how much the box is holding now, .capacity() is the volume of a box, and .max_size() is the maximum size a box could be before it undergoes gravitational collapse.

Example

vector<int> v;
cout << "size="     << v.size()     << '\n'
     << "capacity=" << v.capacity() << '\n'
     << "max_size=" << v.max_size() << '\n';

size=0
capacity=0
max_size=2305843009213693951

vector<char> v(42);
cout << "size="     << v.size()     << '\n'
     << "capacity=" << v.capacity() << '\n'
     << "max_size=" << v.max_size() << '\n'
     << "2⁶³=" << (1ULL<<63) << '\n';

size=42
capacity=42
max_size=9223372036854775807
2⁶³=9223372036854775808

Dynamic Resizing

vector<int> v;
for (int i=0; i<18; i++) {
    v.push_back(42);
    cout << "size="     << v.size()     << ' '
         << "capacity=" << v.capacity() << '\n';
}

size=1 capacity=1
size=2 capacity=2
size=3 capacity=4
size=4 capacity=4
size=5 capacity=8
size=6 capacity=8
size=7 capacity=8
size=8 capacity=8
size=9 capacity=16
size=10 capacity=16
size=11 capacity=16
size=12 capacity=16
size=13 capacity=16
size=14 capacity=16
size=15 capacity=16
size=16 capacity=16
size=17 capacity=32
size=18 capacity=32

Change sizes

• .reserve(n): change the allocated size of this container. However, you can’t allocate less than your current .size(). This changes .capacity(). This does not change .size().
• .resize(n): change the size of this container. If it had more than n elements before, we lose some. If it had fewer than n elements, add default-constructed elements. This changes .size(). It may change .capacity().
• .shrink_to_fit(): Get rid of any additional reserved memory, as if via .reserve(.size()).

Example with .reserve()

vector<int> v;
v.reserve(7);
for (int i=0; i<18; i++) {
    v.push_back(42);
    cout << "size="     << v.size()     << ' '
         << "capacity=" << v.capacity() << '\n';
}

size=1 capacity=7
size=2 capacity=7
size=3 capacity=7
size=4 capacity=7
size=5 capacity=7
size=6 capacity=7
size=7 capacity=7
size=8 capacity=14
size=9 capacity=14
size=10 capacity=14
size=11 capacity=14
size=12 capacity=14
size=13 capacity=14
size=14 capacity=14
size=15 capacity=28
size=16 capacity=28
size=17 capacity=28
size=18 capacity=28

Example with .resize()

vector<int> v;
v.resize(7);
for (int i=0; i<18; i++) {
    v.push_back(42);
    cout << "size="     << v.size()     << ' '
         << "capacity=" << v.capacity() << '\n';
}

size=8 capacity=14
size=9 capacity=14
size=10 capacity=14
size=11 capacity=14
size=12 capacity=14
size=13 capacity=14
size=14 capacity=14
size=15 capacity=28
size=16 capacity=28
size=17 capacity=28
size=18 capacity=28
size=19 capacity=28
size=20 capacity=28
size=21 capacity=28
size=22 capacity=28
size=23 capacity=28
size=24 capacity=28
size=25 capacity=28

Strings, too

string s;
for (int i=0; i<18; i++) {
    s += 'x';
    cout << "size="     << s.size()     << ' '
         << "capacity=" << s.capacity() << '\n';
}

size=1 capacity=15
size=2 capacity=15
size=3 capacity=15
size=4 capacity=15
size=5 capacity=15
size=6 capacity=15
size=7 capacity=15
size=8 capacity=15
size=9 capacity=15
size=10 capacity=15
size=11 capacity=15
size=12 capacity=15
size=13 capacity=15
size=14 capacity=15
size=15 capacity=15
size=16 capacity=30
size=17 capacity=30
size=18 capacity=30

How about a set?

set<int> s;
for (int i=0; i<17; i++) {
    s.insert(i);
    cout << "size="     << s.size()     << ' '
         << "capacity=" << s.capacity() << '\n';
}

c.cc:5: error: ‘class std::set<int>’ has no member named ‘capacity’

• The only reason that vector and string have this .capacity() business is that it’s more efficient to have the data be all together (i.e., contiguous).
• This doesn’t apply to a set or map, which generally use binary trees, or list, which is a doubly-linked list.

size()

There is a free function, size(), which returns the size of a container (by calling .size()) or the size of an array (because the size of an array is known at compile-time).

vector<int> v = {11,22,33};
double d[] = {1.2, 3.4, 5.6, 7.8};
cout << v.size() << ' ' << size(v) << '\n'
     << size(d) << '\n';

3 3
4

It’s used for template code (patience) that has to work for a container or an old-style C array. It’s also darn handy for getting the size of a C array while keeping DRY.

sizeof() is not .size() is not size()

short *p, a[42];
string s;
cout << sizeof(p)              << '\n'
     << sizeof(a)              << '\n'
     << sizeof(a[0])           << '\n'
     << sizeof(a)/sizeof(a[0]) << '\n'
     << size(a)                << '\n'
     << sizeof(s)              << '\n'
     << s.size()               << '\n';

8
84
2
42
42
32
0

There are three functions with similar names:

sizeof()

a compile-time function that returns the size of an object in bytes

.size()

a container method, returns number of elements

size()

a free function, identical to .size() for containers, returns the number of elements for a C array.