Show Lecture.Vector as a slide show.

CS253 Vector

Inclusion

To use vector, you need to:

    
#include <vector>

Old School

• Old-style C arrays still work just fine.
• They have no methods.
• They are NOT stretchy.

int a[10];
for (int i=0; i<10; i++)
    a[i] = i*i;
for (int i=0; i<10; i++)
    cout << a[i] << ' ';

0 1 4 9 16 25 36 49 64 81 

int a[10];
for (int i=0; i<10; i++)
    a[i] = i*i;
for (auto val : a)
    cout << val << ' ';

0 1 4 9 16 25 36 49 64 81 

vector

• A vector is a stretchy array

vector<int> v;
for (int i=0; i<10; i++)
    v.push_back(i*i*i);
for (size_t i=0; i<v.size(); i++)
    cout << v[i] << ' ';

0 1 8 27 64 125 216 343 512 729 

vector<int> v(10);
for (int i=0; i<10; i++)
    v[i] = i*i*i;
for (auto val : v)
    cout << val << ' ';

0 1 8 27 64 125 216 343 512 729 

Useful vector methods

For a vector named v:

• v.size(): returns the current length, as a size_t
• v.push_back(): add an element to the end
• v.pop_back(): remove an element from the end
• v.erase(): remove an element from anywhere
• v[i] or v.at(i): indexing
• and many more methods

vector constructors

Vectors have several constructors, including:

// default size of zero elements
vector<float> a;

// initialize with these values
vector<float> b = {1.2, 3.4};

// resize to this many zeroes
vector<float> c(170);

if (a != b)
    cout << "unequal\n";

unequal

vector of anything

You can have a vector of nearly any type, but you have to pick one.

vector<string> some_strings;
vector<double> some_doubles(42);
vector<bool> some_bools = {true, false};

cout << some_strings.size() << '\n';
cout << some_doubles.size() << '\n';
cout << some_bools.size() << '\n';

0
42
2

vector oops;  // 🦡

c.cc:1: error: class template argument deduction failed:

Stretchy under certain circumstances

Sure, vector is stretchy at certain points. It’s not magic. The vector grows when:

• given a size_t ctor argument
• calling vector::push_back(), vector::emplace_back(), vector::insert()
• calling vector::resize() (not vector::reserve())

vector<string> spices;
cout << spices.size() << endl;
spices[0] = "Posh"; // 🦡 BAD!

0
SIGSEGV: Segmentation fault

Stretchy under certain circumstances

• You can only access [n ] if there are at least n +1 elements already in the vector.
• vector::push_back() guarantees that there’s enough space for another element, and then adds the element at the end (back).

vector<string> spices = {"Posh", "Sporty"};
cout << spices.size() << '\n';
spices.push_back("Scary");
cout << spices.size() << '\n';
spices.push_back("Ginger");
spices.push_back("Baby");
cout << spices.size() << '\n';
for (auto s : spices)
    cout << s << ' ';

2
3
5
Posh Sporty Scary Ginger Baby 

Stretchy under certain circumstances

• If you know the initial size, you can say so in the ctor. This will initialize the vector to that many empty (default-constructed) entries.

vector<string> spices(5);
spices[0] = "Posh";
spices[1] = "Sporty";
spices[2] = "Scary";
cout << spices.size() << '\n';
spices[3] = "Ginger";
spices[4] = "Baby";
cout << spices.size() << '\n';
for (auto s : spices)
    cout << s << ' ';

5
5
Posh Sporty Scary Ginger Baby 

• Don’t combine this with .push_back(), though.

Stretchy under certain circumstances

• Of course, if you know the initial data, just say so without counting.

vector<string> spices = { "Posh", "Sporty", "Scary",
                          "Ginger", "Baby" };
cout << spices.size() << '\n';
for (auto s : spices)
    cout << s << ' ';

5
Posh Sporty Scary Ginger Baby 

CTAD

The C++17 feature Class Template Argument Deduction, or CTAD, allows the omission of the type in some cases when a container is being initialized:

vector winkies = {11,22,33};   // no <int>!
for (auto v : winkies)
    cout << v << ' ';

11 22 33 

vector quadlings{'G', 'l', 'i', 'n', 'd', 'a'};
for (auto v : quadlings)
    cout << v << ' ';

G l i n d a 

CTAD requires initialization

CTAD deduces the type from the initialization, so you’ve got to have initialization for this to work. You can’t have a vector of type whatever-you-put-into-it-later:

vector<int> munchkins;
munchkins.push_back(1900);
cout << munchkins[0];

1900

vector gillikins;  // 🦡
gillikins.push_back(1904);

c.cc:1: error: class template argument deduction failed:

CTAD with other types

CTAD works with all standard C++ containers, and other types that we haven’t discussed, yet:

set<const char *> s = {"Every", "Good", "Boy", "Does", "Fine"};
list<double> l{440.0, 880.0};
pair<set<const char *>, list<double>> p(s,l);
cout << p.first.size() << ' ' << p.second.size();

5 2

set s = {"Every", "Good", "Boy", "Does", "Fine"};
list l{440.0, 880.0};
pair p(s,l);
cout << p.first.size() << ' ' << p.second.size();

5 2