Show Lecture.STLContainers as a slide show.

CS253 STL Containers

Overview

• The most popular containers:
  • vector: variable-length array
  • array: a fixed-length array
  • string: vector<char> and more methods
  • list: doubly-linked list
  • forward_list: singly-linked list
  • set: an ordered binary tree
  • multiset: a set where duplicates are permitted
  • map: a tree of key,value (or a translation)
  • multimap: a map where duplicate keys are permitted
  • deque: a vector with push/pop at either end
  • unordered_set: hash version of set
  • unordered_multiset: hash version of multiset
  • unordered_map: hash version of map
  • unordered_multimap: hash version of multimap

Attributes

Name	Order	Search	Insert	Append
string, vector	no	O(n )	O(n )	O(1)
deque	no	O(n )	O(n )	O(1)
list, forward_list	no	O(n )	O(1)	O(1)
[multi]set	<	O(log n )	O(log n )
[multi]map	<	O(log n )	O(log n )
unordered_[multi]set	hashed	O(1)	O(1)
unordered_[multi]map	hashed	O(1)	O(1)

Popular methods

Universal container methods	Popular container methods
default ctor	ctor(n)
copy ctor	ctor(iter, iter)
.operator=	.size()
.empty()	.insert() / .erase()
.max_size()	.find() / .count()
.clear()	.front() / .back()
.begin() / .end()	.rbegin() / .rend()
.cbegin() / .cend()	.crbegin() / .crend()

Popular types

Universal types	Popular types
value_type	key_type
iterator	reverse_iterator
const_iterator	const_reverse_iterator
size_type

string

string con = "The quick brown fox jumps over a lazy dog.";
cout << "size=" << con.size() << ' ';
for (auto c : con)
    cout << c;

size=42 The quick brown fox jumps over a lazy dog.

• This displays a std::string.
• After this, the same code will be shown, but using different containers.
• Sure, we could have just said cout << con; since we’re using a std::string, but but that won’t work for other container types.

vector

string s = "The quick brown fox jumps over a lazy dog.";
vector<char> con(s.begin(), s.end());
cout << "size=" << con.size() << ' ';
for (auto c : con)
    cout << c;

size=42 The quick brown fox jumps over a lazy dog.

We can’t initialize a vector directly from a C-style string, so we put it into a C++ string, and initialize the vector from that, using the two-iterator constructor.

array

string s = "The quick brown fox jumps over a lazy dog.";
array<char, 42> con;
for (size_t i=0; i<s.size(); i++)
    con[i] = s[i];
cout << "size=" << con.size() << ' ';
for (auto c : con)
    cout << c;

size=42 The quick brown fox jumps over a lazy dog.

• A std::array is not a C array, but its size is fixed at compile time.
• Its storage is all on the stack, with no dynamic memory allocation, so it’s as effecient as a C array. It has typical STL container methods such as .size(), .begin(), .end(), etc.
• It lacks the two-iterator ctor, perhaps due to its fixed size.

list

string s = "The quick brown fox jumps over a lazy dog.";
list<char> con(s.begin(), s.end());
cout << "size=" << con.size() << ' ';
for (auto c : con)
    cout << c;

size=42 The quick brown fox jumps over a lazy dog.

• A list is a doubly-linked list.
• Large storage overhead, but insertion is cheap, and you can easily go forward & backward.
• Unlike a vector, .push_back() never causes reallocation.
• When you iterate over a list, you don’t get a node—those are internal. You just get the value stored in the node.

forward_list

string s = "The quick brown fox jumps over a lazy dog.";
forward_list<char> con(s.begin(), s.end());
for (auto c : con)
    cout << c;

The quick brown fox jumps over a lazy dog.

• A forward_list (formerly slist) is a singly-linked list.
• It has storage overhead, but insertion is cheap, and you can easily go forward.
• When you iterate over a forward_list, you don’t get a node—those are internal. You just get the value stored in the node.
• There is no forward_list::size(). Perhaps this was to keep an empty forward_list small, by not having a size_t to keep the size:

  forward_list<string> fl;
  cout << sizeof(fl);
  

  8

deque

string s = "The quick brown fox jumps over a lazy dog.";
deque<char> con(s.begin(), s.end());
cout << "size=" << con.size() << ' ';
for (auto c : con)
    cout << c;

size=42 The quick brown fox jumps over a lazy dog.

• Pronounced like “deck”: 🃚 🂻 🂭 🃎 🃑
• A deque is a double-ended queue.
• It’s like a vector, but you can easily push or pop from the front or the back.

set

string s = "The quick brown fox jumps over a lazy dog.";
set<char> con(s.begin(), s.end());
cout << "size=" << con.size() << ' ';
for (auto c : con)
    cout << c;

size=28  .Tabcdefghijklmnopqrsuvwxyz

• Generally implemented as a binary tree. Hence, it is sorted, has fairly large storage overhead, and has O(log n ) find/insertion/deletion times.
• It’s intrinsically, always, sorted. You don’t have to sort it. When you .insert() something, it gets put into the right place to preserve the sortyness.
• When you iterate over a set, you don’t get a node—those are internal. You just get the value stored in the node.

multiset

string s = "The quick brown fox jumps over a lazy dog.";
multiset<char> con(s.begin(), s.end());
cout << "size=" << con.size() << ' ';
for (auto c : con)
    cout << c;

size=42         .Taabcdeefghijklmnoooopqrrsuuvwxyz

A multiset is like a set, but allows copies.

unordered_set

string s = "The quick brown fox jumps over a lazy dog.";
unordered_set<char> con(s.begin(), s.end());
cout << "size=" << con.size() << ' ';
for (auto c : con)
    cout << c;

size=28 gdyzlavspmjxfn.huerTqickbo w

• Formerly hash_set
• A hash table implementation that acts like a set (no duplicates allowed).
• When you iterate over a unordered_set, you don’t get a bucket—those are internal. You just get the value stored in the bucket.

unordered_multiset

string s = "The quick brown fox jumps over a lazy dog.";
unordered_multiset<char> con(s.begin(), s.end());
cout << "size=" << con.size() << ' ';
for (auto c : con)
    cout << c;

size=42 .gdyzlaaspmjxfnwoooorrbkciuuqv        eehT

A unordered_multiset (formerly hash_multiset), is a hash table implementation like a multiset.

map, multimap, unordered_map, unordered_multimap

• map, multimap, unordered_map, and unordered_multimap are like
  set, multiset, unordered_set, and unordered_multiset, except that they store pairs.
• You can think of them as look-up tables.
• You can think of them as mapping functions.

map example

map<string, double> gpa = {
    { "Jack",  3.998 },
    { "Russ",  4.20  },
    { "Phil",  3.92  },
    { "Craig", 2.0   }
};

for (auto p : gpa)
    cout << p.first << " has a GPA of " << p.second << '\n';
cout << "Jack’s GPA is: " << gpa["Jack"] << '\n';

Craig has a GPA of 2
Jack has a GPA of 3.998
Phil has a GPA of 3.92
Russ has a GPA of 4.2
Jack’s GPA is: 3.998

• A map<A,B> is like a set<pair<const A,B>>, ordered by A.
• When you iterate over a map, you get a pair, with public data members (😲) .first and .second.