【C++】朝花夕拾——STL vector

STL之vector篇

　　N久之前是拿C的数组实现过vector中的一些简单功能，什么深拷贝、增删查找之类的，以为vector的实现也就是这样了，现在想想真是...too young too naive...ORZ

====================我是分割线=============================

vector属于顺序容器，它的底层实现就是基于array，所以它可以支持随机访问，但是它比array更有效率，因为它动态分配的内存空间。

动态分配的内存空间:

　　　每当vector的capacity==size，并且有新element需要加入时，vector会申请一段大小等于2*capacity的连续内存空间，将原来的数据深拷贝到新的内存地址，然后释放掉原来的内存，并添加新的element到内存中。(这一系列内存操作大大影响了vector的存储效率)

　　    因为，不同的环境下，vector的实现方式有所出入，所以初始化时的capacity大小不一样，但是capacity一定会大于所需的内存大小，预留一定的空间，已减少内存操作。

以下是简单的测试:

环境如下:

g++ 4.8.4

ubuntu 14.04

test1:

#include <iostream>
#include <vector>

using namespace std;

int main() {
    vector<int> v1;
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;
    cout << "v1 max_size = " << v1.max_size() << endl;
    vector<char> v2;
    cout << "v2 size = " << v2.size() << endl;
    cout << "v2 capacity = " << v2.capacity() << endl;
    cout << "v2 max_size = " << v2.max_size() << endl;
    return 0;
}

结果如下: 

v1 size = 0
v1 capacity = 0
v1 max_size = 4611686018427387903
v2 size = 0
v2 capacity = 0
v2 max_size = 18446744073709551615

max_size返回vector可以保持的element个数的上限，这个上限与系统的底层实现有关，不一定可以达到(当内存不够时，不能继续分配）

这里可以看出，此时初始化的capacity==0

test2:

#include <iostream>
#include <vector>

using namespace std;

int main() {
    vector<int> v1;
    for (int i = 0; i < 1; ++i)v1.push_back(i);
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;
    cout << "v1 max_size = " << v1.max_size() << endl;
    vector<char> v2;
    for (int i = 0; i < 1; ++i)v2.push_back(char(i));
    cout << "v2 size = " << v2.size() << endl;
    cout << "v2 capacity = " << v2.capacity() << endl;
    cout << "v2 max_size = " << v2.max_size() << endl;
    return 0;
}

结果如下:

v1 size = 1
v1 capacity = 1
v1 max_size = 4611686018427387903
v2 size = 1
v2 capacity = 1
v2 max_size = 18446744073709551615

此时，插入一个元素，capacity==1，size也为1

test3:

#include <iostream>
#include <vector>

using namespace std;

int main() {
    vector<int> v1;
    for (int i = 0; i < 2; ++i)v1.push_back(i);
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;
    cout << "v1 max_size = " << v1.max_size() << endl;
    vector<char> v2;
    for (int i = 0; i < 2; ++i)v2.push_back(char(i));
    cout << "v2 size = " << v2.size() << endl;
    cout << "v2 capacity = " << v2.capacity() << endl;
    cout << "v2 max_size = " << v2.max_size() << endl;
    return 0;
}

结果如下:

v1 size = 2
v1 capacity = 2
v1 max_size = 4611686018427387903
v2 size = 2
v2 capacity = 2
v2 max_size = 18446744073709551615

当插入两个元素时，capacity = 2 * capacity(即为2），然后拷贝之前的数据到新的内存空间，添加新元素，则size==2

test4:

#include <iostream>
#include <vector>

using namespace std;

int main() {
    vector<int> v1;
    for (int i = 0; i < 3; ++i)v1.push_back(i);
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;
    cout << "v1 max_size = " << v1.max_size() << endl;
    vector<char> v2;
    for (int i = 0; i < 3; ++i)v2.push_back(char(i));
    cout << "v2 size = " << v2.size() << endl;
    cout << "v2 capacity = " << v2.capacity() << endl;
    cout << "v2 max_size = " << v2.max_size() << endl;
    return 0;
}

结果如下:

v1 size = 3
v1 capacity = 4
v1 max_size = 4611686018427387903
v2 size = 3
v2 capacity = 4
v2 max_size = 18446744073709551615

又接着添加一个新元素，capacity翻倍(变成4），size增加1，即3

添加新元素的原理get  >   <

===========================第二分割线===================================

以下为删除原理

test5:

#include <iostream>
#include <vector>

using namespace std;

int main() {
    vector<int> v1;
    for (int i = 0; i < 129; ++i)v1.push_back(i);
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;
    cout << "v1 max_size = " << v1.max_size() << endl;
    vector<char> v2;
    for (int i = 0; i < 129; ++i)v2.push_back(char(i));
    cout << "v2 size = " << v2.size() << endl;
    cout << "v2 capacity = " << v2.capacity() << endl;
    cout << "v2 max_size = " << v2.max_size() << endl;
    return 0;
}

结果如下:

v1 size = 129
v1 capacity = 256
v1 max_size = 4611686018427387903
v2 size = 129
v2 capacity = 256
v2 max_size = 18446744073709551615

test6:

#include <iostream>
#include <vector>

using namespace std;

int main() {
    vector<int> v1;
    for (int i = 0; i < 129; ++i)v1.push_back(i);
    v1.pop_back();
    v1.pop_back();
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;
    cout << "v1 max_size = " << v1.max_size() << endl;
    vector<char> v2;
    for (int i = 0; i < 129; ++i)v2.push_back(char(i));
    v2.pop_back();
    v2.pop_back();
    cout << "v2 size = " << v2.size() << endl;
    cout << "v2 capacity = " << v2.capacity() << endl;
    cout << "v2 max_size = " << v2.max_size() << endl;
    return 0;
}

结果如下:

v1 size = 127
v1 capacity = 256
v1 max_size = 4611686018427387903
v2 size = 127
v2 capacity = 256
v2 max_size = 18446744073709551615

(⊙v⊙)嗯，看来内存的动态改变并不会随着size的减小的缩减capacity，避免了不必要的内存操作。

C++手册中也有说明，pop_back的操作只是size减1

=============================第三分割线==================================

接着，是各个函数的底层理解

assign() 

　　有三种形参形式(iterator begin，iterator end)初始化为迭代器指向的容器之间的内容；(n, value)初始化N个值为value的元素；(initializer_list<value_type> il)深拷贝

back()

　　返回vector中最后一个值，当vector为空时，调用该函数会报错

front()

　　返回vector中第一个值，当vector为空时，调用该函数会报错

begin()

　　返回指向第一个element的迭代器

end()

　　返回指向past-the-end element的迭代器，past-the-end element是理论上应该跟在最后一个element后面的位置，但是那个位置上并没有存放element

cbegin()

　　返回一个首个元素的const_iterator迭代器，用该迭代器访问元素是是不可以用来改变元素本身值的(即便元素本身并非const类型）

cend()

　　返回一个past-the-last element的const_iterator迭代器

crbegin()

　　r表示逆序，c表示const，此处返回vector中逆序第一个element的const_iterator

crend()

　　此处返回vector中第一个element的const_iterator，功能与cbegin()相同

rbegin()

　　返回逆序第一个element的迭代器，即为指向end()的前一个位置的迭代器

rend()

　　返回逆序的最后一个element的迭代器，即指向第一个element

capacity()

　　返回vector容器已开辟的内存可存放element的个数

size()

　　返回当前vector中已有的element个数，size <= capacity

max_size()

　　返回一个理论上允许的capacity上限值，但不一定能达到

data()

　　返回一个指向当前容器内存起始地址的指针，由于vector的内存分配是连续的，所以可以直接用指针offset来为容器中的element赋值

　　std::vector<int> myvector (5);

　　int* p = myvector.data();

　　*p = 10;

　　++p;

　　*p = 20;

    p[2] = 100;

emplace()

 关于emplace和insert的区别，如下:

(emplace会调用类的构造函数,代码搬运自:http://stackoverflow.com/questions/14788261/c-stdvector-emplace-vs-insert）

struct Foo
{
  Foo(int n, double x);
};

std::vector<Foo> v;
v.emplace(someIterator, 42, 3.1416);
v.insert(someIterator, Foo(42, 3.1416));

emplace_back()

　　调用构造函数，然后把element加到末尾

insert()

　　可以把element插入到迭代器指定的element之前的位置。

　　由于底层实现时array，除了在末尾插入element，在其他的位置插入，都需要将指定插入位置后面的element移位，这将导致效率低下

　　如果插入之后size>capacity，则参考push_back的重新申请内存空间

erase()

　　删除指定区间或者位置的element，同样，除了删除末尾的element，删除其他位置的element都需要把后面的element往前移动，也是导致效率低的原因

clear()

　　把所有element都移除，size=0，此时capacity不一定为0(不一定会有释放内存的操作）//正确的释放内存姿势——swap()

empty()

　　size >= 0, 返回false；否则返回true

get_allocator()

　　返回该容器的allocator

operator=

　　拷贝，STL中的拷贝都是深拷贝

operator[]

　　访问节点

push_back()

　　在末尾增加一个element，当size超出时，capacity翻倍(具体看此博文开头)，size++

pop_back()

　　移除末尾一个element, size--

reserve()

　　v.reserve（n）;表示此时V的capacity至少要大于等于300，若capacity<300，则重新申请内存；反之则不用做其他操作；

　　

#include <iostream>
#include <vector>

using namespace std;

int main() {
    vector<int> v1;
    for (int i = 0; i < 129; ++i)v1.push_back(i);
    v1.pop_back();
    v1.pop_back();
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;
    cout << "v1 max_size = " << v1.max_size() << endl;
    vector<char> v2;
    for (int i = 0; i < 129; ++i)v2.push_back(char(i));
    v2.reserve(300);
    cout << "v2 size = " << v2.size() << endl;
    cout << "v2 capacity = " << v2.capacity() << endl;
    cout << "v2 max_size = " << v2.max_size() << endl;
    return 0;
}

结果为:

v1 size = 127
v1 capacity = 256
v1 max_size = 4611686018427387903
v2 size = 129
v2 capacity = 300
v2 max_size = 18446744073709551615

resize()

　　v.resize（n）将会把v中的size调整为n大小，若n > size，则用0补足element；若n < size，则取容器中的前n个element，其与元素移除；

　　v.resize（n，m）将会把v中的size调整为n大小，若n > size，则用m补足element

　　

#include <iostream>
#include <vector>

using namespace std;

int main() {
    vector<int> v1;
    for (int i = 0; i < 10; ++i)v1.push_back(i);
    for (int i = 0; i < v1.size(); i ++)cout << v1[i] << " ";
    cout << endl;
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;
    cout << "v1 max_size = " << v1.max_size() << endl;

    v1.resize(7);
    for (int i = 0; i < v1.size(); i ++)cout << v1[i] << " ";
    cout << endl;
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;
    cout << "v1 max_size = " << v1.max_size() << endl;

    v1.resize(20, 200);
    for (int i = 0; i < v1.size(); i ++)cout << v1[i] << " ";
    cout << endl;
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;
    cout << "v1 max_size = " << v1.max_size() << endl;

    v1.resize(50);
    for (int i = 0; i < v1.size(); i ++)cout << v1[i] << " ";
    cout << endl;
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;
    cout << "v1 max_size = " << v1.max_size() << endl;
    return 0;
}

结果如下:

0 1 2 3 4 5 6 7 8 9
v1 size = 10
v1 capacity = 16
v1 max_size = 4611686018427387903
0 1 2 3 4 5 6
v1 size = 7
v1 capacity = 16
v1 max_size = 4611686018427387903
0 1 2 3 4 5 6 200 200 200 200 200 200 200 200 200 200 200 200 200
v1 size = 20
v1 capacity = 20
v1 max_size = 4611686018427387903
0 1 2 3 4 5 6 200 200 200 200 200 200 200 200 200 200 200 200 200 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
v1 size = 50
v1 capacity = 50
v1 max_size = 4611686018427387903

shrink_to_fit()

　　一般来说capacity满足，capacity>=size；v.shrink_to_fit()操作可以使得，capacity==size，这也是释放内存的操作之一!！!

　　

#include <iostream>
#include <vector>

using namespace std;

int main() {
    vector<int> v1;
    for (int i = 0; i < 10; ++i)v1.push_back(i);
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;
    cout << "v1 max_size = " << v1.max_size() << endl;

    v1.resize(7);
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;
    cout << "v1 max_size = " << v1.max_size() << endl;

    v1.shrink_to_fit();
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;
    cout << "v1 max_size = " << v1.max_size() << endl;

    return 0;
}

结果如下:

v1 size = 10
v1 capacity = 16
v1 max_size = 4611686018427387903
v1 size = 7
v1 capacity = 16
v1 max_size = 4611686018427387903
v1 size = 7
v1 capacity = 7
v1 max_size = 4611686018427387903

swap()

　　v.swap(v2)的操作可以将v，和v2中的element交换

#include <iostream>
#include <vector>

using namespace std;

int main() {
    vector<int> v1;
    for (int i = 0; i < 10; ++i)v1.push_back(i);
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;
    cout << "v1 max_size = " << v1.max_size() << endl;

    vector<int> v2;
    for (int i = 0; i < 20; ++i)v2.push_back(i);
    cout << "v2 size = " << v2.size() << endl;
    cout << "v2 capacity = " << v2.capacity() << endl;
    cout << "v2 max_size = " << v2.max_size() << endl;

    v1.swap(v2);
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;
    cout << "v1 max_size = " << v1.max_size() << endl;
    cout << "v2 size = " << v2.size() << endl;
    cout << "v2 capacity = " << v2.capacity() << endl;
    cout << "v2 max_size = " << v2.max_size() << endl;

    return 0;
}

结果如下:

v1 size = 10
v1 capacity = 16
v1 max_size = 4611686018427387903
v2 size = 20
v2 capacity = 32
v2 max_size = 4611686018427387903
v1 size = 20
v1 capacity = 32
v1 max_size = 4611686018427387903
v2 size = 10
v2 capacity = 16
v2 max_size = 4611686018427387903

小结:

vector是不会自动释放内存的，但是如果size的极大值很大的话会造成capacity很大，浪费内存。

所以以下有几种方法可以释放vector的内存:

　　①shrink_to_fit

#include <iostream>
#include <vector>

using namespace std;

int main() {
    vector<int> v1;
    for (int i = 0; i < 20; ++i)v1.push_back(i);
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;

    v1.clear();
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;

    v1.shrink_to_fit();
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;
    

    return 0;
}

结果:

v1 size = 20

v1 capacity = 32

v1 size = 0
v1 capacity = 32
v1 size = 0
v1 capacity = 0

　　

　　②swap

　　

#include <iostream>
#include <vector>

using namespace std;

int main() {
    vector<int> v1;
    for (int i = 0; i < 20; ++i)v1.push_back(i);
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;

    vector<int>().swap(v1);
    cout << "v1 size = " << v1.size() << endl;
    cout << "v1 capacity = " << v1.capacity() << endl;

    return 0;
}

结果如下:

v1 size = 20
v1 capacity = 32
v1 size = 0
v1 capacity = 0

 

　　　　③指针释放

　　　　

#include <iostream>
#include <vector>

using namespace std;

int main() {
    vector<int> v;
    vector<int>* v1 = &v;
    for (int i = 0; i < 20; ++i)v1->push_back(i);
    cout << "v size = " << v.size() << endl;
    cout << "v capacity = " << v.capacity() << endl;

    delete v1;
    return 0;
}

========================我是第四分割线============================

C++11的标准中，vector新增加的成员函数:

crbegin

crend

cbegin

cend

emplace

emplace_back

data

shrink_to_fit

========================我是第五分割线============================

最后简单说说，高效索引之bitset和vector<bool>

bitset　　可以进行位操作，但是大小固定

vector<bool>   属于vector的一种特殊形式，用bool类型的allocator构造，但是在实现的时候优化了，所以它的element类型并不是bool，而是bit