SET的完整用法
SET的第一行模板原始碼是這樣的:
template<typename _Key, typename _Compare = std::less<_Key>,
typename _Alloc = std::allocator<_Key> >
根據template和English的有關知識,我們可以知道set的定義方式應該是:
set<型別,比較函式(不寫時預設為less<型別>,從小到大),空間配置器(從應用STL角度來說完全可以忽略,實際上也完全一定要忽略不寫)>變數名
第110行:
typedef _Rb_tree<key_type, value_type, _Identity<value_type>, key_compare, _Key_alloc_type> _Rep_type; _Rep_type _M_t; // Red-black tree representing set.
現在我們知道set是用紅黑樹來維護的。
按順序 第149行:
/** * @brief Builds a %set from a range. * @param first An input iterator. * @param last An input iterator. * * Create a %set consisting of copies of the elements from [first,last). * This is linear in N if the range is already sorted, and NlogN * otherwise (where N is distance(first,last)). */ template<typename _InputIterator> set(_InputIterator __first, _InputIterator __last) : _M_t() { _M_t._M_insert_unique(__first, __last); } /** * @brief Builds a %set from a range. * @param first An input iterator. * @param last An input iterator. * @param comp A comparison functor. * @param a An allocator object. * * Create a %set consisting of copies of the elements from [first,last). * This is linear in N if the range is already sorted, and NlogN * otherwise (where N is distance(first,last)). */ template<typename _InputIterator> set(_InputIterator __first, _InputIterator __last, const _Compare& __comp, const allocator_type& __a = allocator_type()) : _M_t(__comp, __a) { _M_t._M_insert_unique(__first, __last); }
這個的意思是可以用迭代器初始化set(不懂可以看下面的測試程式碼),並且速度在資料有序的情況下為O(n)(常數偏大),無序的情況下為O(nlogn)(常數偏小)。
我的測試程式碼:
#include<bits/stdc++.h> using namespace std; vector<int>vec; set<int>st; int main() { for(int i=1;i<=2000000;i++) vec.push_back(i); int t1 = clock(); st = set<int>(vec.begin(),vec.end()); printf("%d\n",clock()-t1); random_shuffle(vec.begin(),vec.end()); t1=clock(); st = set<int>(vec.begin(),vec.end()); printf("%d\n",clock()-t1); }
輸出(不開O2):
1083
3007
輸出(開O2):
724
1689
可以發現在開O2的情況下還是可以接受的,
100000個數只需要100ms左右。
219行
/**
* @brief %Set assignment operator.
* @param x A %set of identical element and allocator types.
*
* All the elements of @a x are copied, but unlike the copy constructor,
* the allocator object is not copied.
*/
set&
operator=(const set& __x)
{
_M_t = __x._M_t;
return *this;
}
set的等於符號過載,O(n)
測試程式碼:
#include<bits/stdc++.h>
using namespace std;
vector<int>vec;
set<int>st;
int main()
{
for(int i=1;i<=100000;i++)
vec.push_back(i);
int t1 = clock();
st = set<int>(vec.begin(),vec.end());
printf("%d\n",clock()-t1);
t1=clock();
set<int>st2 = st;
printf("%d\n",clock()-t1);
}
輸出 :
57
18
689行:
template<typename _Key, typename _Compare, typename _Alloc>
inline void
swap(set<_Key, _Compare, _Alloc>& __x, set<_Key, _Compare, _Alloc>& __y)
{ __x.swap(__y); }
這個swap好像是直接交換指標的,O(1)
測試程式碼:
#include<bits/stdc++.h>
using namespace std;
vector<int>vec;
set<int>st;
int main()
{
for(int i=1;i<=100000;i++)
vec.push_back(i);
int t1 = clock();
st = set<int>(vec.begin(),vec.end());
printf("%d\n",clock()-t1);
t1=clock();
set<int>st2;
st2.swap(st);
printf("%d\n",*st2.begin());
printf("%d\n",clock()-t1);
}
輸出:
57
1
0
943行:
/**
* Erases all the elements. Note that this function only erases the
* elements, and that if the elements themselves are pointers, the
* pointed-to memory is not touched in any way. Managing the pointer is
* the user's responsibility.
*/
void
clear()
{ _M_erase_at_end(this->_M_impl._M_start); }
clear()函式,O(n)的。(Managing the pointer is the user’s responsibility.這話好絕,再也不用set裝指標了)
測試程式碼:
#include<bits/stdc++.h>
using namespace std;
vector<int>vec;
set<int>st;
int main()
{
for(int i=1;i<=100000;i++)
vec.push_back(i);
int t1 = clock();
st = set<int>(vec.begin(),vec.end());
printf("%d\n",clock()-t1);
t1=clock();
st.clear();
printf("%d\n",clock()-t1);
}
輸出:
53
10
272行:
/// Returns the comparison object with which the %set was constructed.
key_compare
key_comp() const
{ return _M_t.key_comp(); }
現在來解釋一下比較函式。
因為STL比較強大,比較函式並不是一個函式,而是一個結構體,裡面過載了()符號,用(a,b)代表比較a和b,並且返回一個值,1代表a<b,0otherwise(被STL註釋語言同化了)
具體看下面程式碼應該能夠有一些瞭解,需要一些template和結構體的前置知識。
測試程式碼:
#include<bits/stdc++.h>
using namespace std;
vector<int>vec;
set<int>st;
struct cmp
{
bool operator ()(int u,int v){ return u<v; }
}a;
int main()
{
printf("%d\n",a(2,1));
printf("%d\n",a(1,2));
less<int>b;
printf("%d\n",b(1,2));
greater<int>c;
printf("%d\n",c(1,2));
}
輸出:
0
1
1
0
285行,8種迭代器:
/**
* Returns a read-only (constant) iterator that points to the first
* element in the %set. Iteration is done in ascending order according
* to the keys.
*/
iterator
begin() const
{ return _M_t.begin(); }
/**
* Returns a read-only (constant) iterator that points one past the last
* element in the %set. Iteration is done in ascending order according
* to the keys.
*/
iterator
end() const
{ return _M_t.end(); }
/**
* Returns a read-only (constant) iterator that points to the last
* element in the %set. Iteration is done in descending order according
* to the keys.
*/
reverse_iterator
rbegin() const
{ return _M_t.rbegin(); }
/**
* Returns a read-only (constant) reverse iterator that points to the
* last pair in the %set. Iteration is done in descending order
* according to the keys.
*/
reverse_iterator
rend() const
{ return _M_t.rend(); }
#ifdef __GXX_EXPERIMENTAL_CXX0X__
/**
* Returns a read-only (constant) iterator that points to the first
* element in the %set. Iteration is done in ascending order according
* to the keys.
*/
iterator
cbegin() const
{ return _M_t.begin(); }
/**
* Returns a read-only (constant) iterator that points one past the last
* element in the %set. Iteration is done in ascending order according
* to the keys.
*/
iterator
cend() const
{ return _M_t.end(); }
/**
* Returns a read-only (constant) iterator that points to the last
* element in the %set. Iteration is done in descending order according
* to the keys.
*/
reverse_iterator
crbegin() const
{ return _M_t.rbegin(); }
/**
* Returns a read-only (constant) reverse iterator that points to the
* last pair in the %set. Iteration is done in descending order
* according to the keys.
*/
reverse_iterator
crend() const
{ return _M_t.rend(); }
好像前面加不加c都一樣。。。迭代器就是一個變數只支援++和–,即只能迭代到下一個或上一個,因為這像迭代的過程,故名迭代器。
begin()是第一個,end()是最後一個的後一個(即是空),體現了STL左閉右開的風格。
rbegin()是最後一個,rend()是最前面的前面一個(即是空),體現了STL左閉右開的風格。
359行:
/// Returns true if the %set is empty.
bool
empty() const
{ return _M_t.empty(); }
/// Returns the size of the %set.
size_type
size() const
{ return _M_t.size(); }
size和empty,都是O(1)的。
#include<bits/stdc++.h>
using namespace std;
vector<int>vec;
set<int>st;
int main()
{
for(int i=1;i<=1000000;i++) st.insert(i);
int t1 = clock();
printf("%d\n",st.size());
printf("%d\n",st.empty());
printf("%d\n",clock()-t1);
}
輸出:
1000000
0
0
393行
// insert/erase
/**
* @brief Attempts to insert an element into the %set.
* @param x Element to be inserted.
* @return A pair, of which the first element is an iterator that points
* to the possibly inserted element, and the second is a bool
* that is true if the element was actually inserted.
*
* This function attempts to insert an element into the %set. A %set
* relies on unique keys and thus an element is only inserted if it is
* not already present in the %set.
*
* Insertion requires logarithmic time.
*/
std::pair<iterator, bool>
insert(const value_type& __x)
{
std::pair<typename _Rep_type::iterator, bool> __p =
_M_t._M_insert_unique(__x);
return std::pair<iterator, bool>(__p.first, __p.second);
}
insert是有返回值的!!!Insertion requires logarithmic time。。。
看註釋應該就懂了,相同元素不重複插入,pair<set::iterater , bool>的前一個是插入的元素的迭代器,後一個是插入是否成功。
測試程式碼
#include<bits/stdc++.h>
using namespace std;
vector<int>vec;
set<int>st;
int main()
{
pair<set<int>::iterator , bool>tmp = st.insert(1);
printf("%d %d\n",*tmp.first,tmp.second);
pair<set<int>::iterator , bool>res = st.insert(1);
printf("%d %d\n",*res.first,res.second);
pair<set<int>::iterator , bool>res1 = st.insert(2);
printf("%d %d\n",*res1.first,res1.second);
}
輸出:
1 1
1 0
2 1
415行:
/**
* @brief Attempts to insert an element into the %set.
* @param position An iterator that serves as a hint as to where the
* element should be inserted.
* @param x Element to be inserted.
* @return An iterator that points to the element with key of @a x (may
* or may not be the element passed in).
*
* This function is not concerned about whether the insertion took place,
* and thus does not return a boolean like the single-argument insert()
* does. Note that the first parameter is only a hint and can
* potentially improve the performance of the insertion process. A bad
* hint would cause no gains in efficiency.
*
* For more on "hinting", see:
* http://gcc.gnu.org/onlinedocs/libstdc++/manual/bk01pt07ch17.html
*
* Insertion requires logarithmic time (if the hint is not taken).
*/
iterator
insert(iterator __position, const value_type& __x)
{ return _M_t._M_insert_unique_(__position, __x); }
指定位置的insert,能夠potentially improve the performance of the insertion process.(卡常必備???)
測試程式碼:
#include<bits/stdc++.h>
using namespace std;
vector<int>vec;
set<int>st;
set<int>::iterator it;
int main()
{
st.insert(1);
int t1=clock();
it = st.begin();
for(int i=2;i<=100000;i++)
{
it++;
st.insert(it,i);
/*
st.insert(i);
*/
}
printf("%d\n",clock()-t1);
printf("%d\n",st.size());
st.clear();
st.insert(1);
t1=clock();
it = st.begin();
for(int i=2;i<=100000;i++)
{
/*
it++;
st.insert(it,i);
*/
st.insert(i);
}
printf("%d\n",clock()-t1);
printf("%d\n",st.size());
}
輸出:
30
100000
74
100000
效果好像還不錯,vector初始化好像要55ms來著。。。
465行:
/**
* @brief Erases an element from a %set.
* @param position An iterator pointing to the element to be erased.
*
* This function erases an element, pointed to by the given iterator,
* from a %set. Note that this function only erases the element, and
* that if the element is itself a pointer, the pointed-to memory is not
* touched in any way. Managing the pointer is the user's responsibility.
*/
void
erase(iterator __position)
{ _M_t.erase(__position); }
/**
* @brief Erases elements according to the provided key.
* @param x Key of element to be erased.
* @return The number of elements erased.
*
* This function erases all the elements located by the given key from
* a %set.
* Note that this function only erases the element, and that if
* the element is itself a pointer, the pointed-to memory is not touched
* in any way. Managing the pointer is the user's responsibility.
*/
size_type
erase(const key_type& __x)
{ return _M_t.erase(__x); }
/**
* @brief Erases a [first,last) range of elements from a %set.
* @param first Iterator pointing to the start of the range to be
* erased.
* @param last Iterator pointing to the end of the range to be erased.
*
* This function erases a sequence of elements from a %set.
* Note that this function only erases the element, and that if
* the element is itself a pointer, the pointed-to memory is not touched
* in any way. Managing the pointer is the user's responsibility.
*/
void
erase(iterator __first, iterator __last)
{ _M_t.erase(__first, __last); }
用迭代器無返回值,用值刪除會返回刪除的個數(意味著用multiset的會將同一個值的刪完),迭代器區間刪除,作者也懶得分析複雜度了。
測試程式碼:
#include<bits/stdc++.h>
using namespace std;
vector<int>vec;
set<int>st;
set<int>::iterator it;
int main()
{
st.insert(1);
it = st.begin();
for(int i=2;i<=1000000;i++)
{
it++;
st.insert(it,i);
}
int t1=clock();
printf("%d\n",st.erase(233));
printf("%d\n",st.erase(233));
for(int i=1;i<=100000;i++)
st.erase(i);
printf("%d\n",clock()-t1);
}
輸出:
1
0
67
刪除常數也可以接受吧。
518行:
// set operations:
/**
* @brief Finds the number of elements.
* @param x Element to located.
* @return Number of elements with specified key.
*
* This function only makes sense for multisets; for set the result will
* either be 0 (not present) or 1 (present).
*/
size_type
count(const key_type& __x) const
{ return _M_t.find(__x) == _M_t.end() ? 0 : 1; }
這程式碼好露骨啊。
535行:
/**
* @brief Tries to locate an element in a %set.
* @param x Element to be located.
* @return Iterator pointing to sought-after element, or end() if not
* found.
*
* This function takes a key and tries to locate the element with which
* the key matches. If successful the function returns an iterator
* pointing to the sought after element. If unsuccessful it returns the
* past-the-end ( @c end() ) iterator.
*/
iterator
find(const key_type& __x)
{ return _M_t.find(__x); }
const_iterator
find(const key_type& __x) const
{ return _M_t.find(__x); }
find函式,返回迭代器,沒有就返回end();
關於後面那個函式我的理解是,set如果是const的就返回const_iterator,歡迎評論。
557行:
/**
* @brief Finds the beginning of a subsequence matching given key.
* @param x Key to be located.
* @return Iterator pointing to first element equal to or greater
* than key, or end().
*
* This function returns the first element of a subsequence of elements
* that matches the given key. If unsuccessful it returns an iterator
* pointing to the first element that has a greater value than given key
* or end() if no such element exists.
*/
iterator
lower_bound(const key_type& __x)
{ return _M_t.lower_bound(__x); }
const_iterator
lower_bound(const key_type& __x) const
{ return _M_t.lower_bound(__x); }
//@}
//@{
/**
* @brief Finds the end of a subsequence matching given key.
* @param x Key to be located.
* @return Iterator pointing to the first element
* greater than key, or end().
*/
iterator
upper_bound(const key_type& __x)
{ return _M_t.upper_bound(__x); }
const_iterator
upper_bound(const key_type& __x) const
{ return _M_t.upper_bound(__x); }
lower_bound()和upper_bound(),意義顯然。
594行:
* @brief Finds a subsequence matching given key.
* @param x Key to be located.
* @return Pair of iterators that possibly points to the subsequence
* matching given key.
*
* This function is equivalent to
* @code
* std::make_pair(c.lower_bound(val),
* c.upper_bound(val))
* @endcode
* (but is faster than making the calls separately).
*
* This function probably only makes sense for multisets.
*/
std::pair<iterator, iterator>
equal_range(const key_type& __x)
{ return _M_t.equal_range(__x); }
std::pair<const_iterator, const_iterator>
equal_range(const key_type& __x) const
{ return _M_t.equal_range(__x); }
一次計算出lower_bound(x)和upper_bound(x)況且只用一次的時間。。。。好像很有用(嗎?)。equal_range…
627行:
/**
* @brief Set equality comparison.
* @param x A %set.
* @param y A %set of the same type as @a x.
* @return True iff the size and elements of the sets are equal.
*
* This is an equivalence relation. It is linear in the size of the sets.
* Sets are considered equivalent if their sizes are equal, and if
* corresponding elements compare equal.
*/
template<typename _Key, typename _Compare, typename _Alloc>
inline bool
operator==(const set<_Key, _Compare, _Alloc>& __x,
const set<_Key, _Compare, _Alloc>& __y)
{ return __x._M_t == __y._M_t; }
/**
* @brief Set ordering relation.
* @param x A %set.
* @param y A %set of the same type as @a x.
* @return True iff @a x is lexicographically less than @a y.
*
* This is a total ordering relation. It is linear in the size of the
* maps. The elements must be comparable with @c <.
*
* See std::lexicographical_compare() for how the determination is made.
*/
template<typename _Key, typename _Compare, typename _Alloc>
inline bool
operator<(const set<_Key, _Compare, _Alloc>& __x,
const set<_Key, _Compare, _Alloc>& __y)
{ return __x._M_t < __y._M_t; }
set的大小比較,<號被定義為字典序???線性複雜度。
multiset就是支援重複元素而已,記得erase的區別,很致命。。。。
完