C++ STL源碼學習(之hash_table篇)

stl_hash_table.h

這不屬於C++標準,是SGI STL標準的一部分,用於輔助實現hash_map和hash_set

/// Hashtable class, used to implement the hashed associative containers
/// hash_set, hash_map, hash_multiset, and hash_multimap.

///STL HashTable採用的是所謂的開鏈哈希法,依靠一個類似vector<list<T>>來實現.
///先通過哈希函數確定所需處理元素應當在vector中的那個位置(可以一步找到),vector
///中的每個元素我們稱之爲一個桶,則這個過程即是尋找桶的過程,每個桶實際是一個
///list<T>元素,若要處理的元素存在,則必然在這個桶之中,然後順序遍歷這個list即
///可確定該元素是否存在或者所在位置.由於第一步的哈希使得桶中存放的元素一
///般很少,因此遍歷查詢過程比較高效.

template <class _Val>
struct _Hashtable_node
{
  _Hashtable_node* _M_next;
  _Val _M_val;
};

///val:存儲的值
///key:對應的鍵
///_HashFcn:所採用的hash函數類型
///_ExtractKey:用於從存儲對象值中抽出鍵對象的函數,當hash_table中
///存儲pair類型實現hash_map時最有用.
///_EqualKey:用於確定兩個鍵值是否相等的函數
template <class _Val, class _Key, class _HashFcn,
          class _ExtractKey, class _EqualKey, class _Alloc = alloc>
class hashtable;

template <class _Val, class _Key, class _HashFcn,
          class _ExtractKey, class _EqualKey, class _Alloc>
struct _Hashtable_iterator;

template <class _Val, class _Key, class _HashFcn,
          class _ExtractKey, class _EqualKey, class _Alloc>
struct _Hashtable_const_iterator;

template <class _Val, class _Key, class _HashFcn,
          class _ExtractKey, class _EqualKey, class _Alloc>
struct _Hashtable_iterator {
  typedef hashtable<_Val,_Key,_HashFcn,_ExtractKey,_EqualKey,_Alloc>
          _Hashtable;
  typedef _Hashtable_iterator<_Val, _Key, _HashFcn,
                              _ExtractKey, _EqualKey, _Alloc>
          iterator;
  typedef _Hashtable_const_iterator<_Val, _Key, _HashFcn,
                                    _ExtractKey, _EqualKey, _Alloc>
          const_iterator;

  typedef _Hashtable_node<_Val> _Node;

  typedef forward_iterator_tag iterator_category;
  typedef _Val value_type;
  typedef ptrdiff_t difference_type;
  typedef size_t size_type;
  typedef _Val& reference;
  typedef _Val* pointer;

  _Node* _M_cur;
  _Hashtable* _M_ht;

  _Hashtable_iterator(_Node* __n, _Hashtable* __tab)
    : _M_cur(__n), _M_ht(__tab) {}

  _Hashtable_iterator() {}
  reference operator*() const { return _M_cur->_M_val; }

  pointer operator->() const { return &(operator*()); }

  iterator& operator++();
  iterator operator++(int);
  bool operator==(const iterator& __it) const
    { return _M_cur == __it._M_cur; }
  bool operator!=(const iterator& __it) const
    { return _M_cur != __it._M_cur; }
};


template <class _Val, class _Key, class _HashFcn,
          class _ExtractKey, class _EqualKey, class _Alloc>
struct _Hashtable_const_iterator {
  typedef hashtable<_Val,_Key,_HashFcn,_ExtractKey,_EqualKey,_Alloc>
          _Hashtable;
  typedef _Hashtable_iterator<_Val,_Key,_HashFcn,
                              _ExtractKey,_EqualKey,_Alloc>
          iterator;
  typedef _Hashtable_const_iterator<_Val, _Key, _HashFcn,
                                    _ExtractKey, _EqualKey, _Alloc>
          const_iterator;
  typedef _Hashtable_node<_Val> _Node;

  typedef forward_iterator_tag iterator_category;
  typedef _Val value_type;
  typedef ptrdiff_t difference_type;
  typedef size_t size_type;
  typedef const _Val& reference;
  typedef const _Val* pointer;

  const _Node* _M_cur;
  const _Hashtable* _M_ht;

  _Hashtable_const_iterator(const _Node* __n, const _Hashtable* __tab)
    : _M_cur(__n), _M_ht(__tab) {}
  _Hashtable_const_iterator() {}
  _Hashtable_const_iterator(const iterator& __it)
    : _M_cur(__it._M_cur), _M_ht(__it._M_ht) {}
  reference operator*() const { return _M_cur->_M_val; }
  pointer operator->() const { return &(operator*()); }
  const_iterator& operator++();
  const_iterator operator++(int);
  bool operator==(const const_iterator& __it) const
    { return _M_cur == __it._M_cur; }
  bool operator!=(const const_iterator& __it) const
    { return _M_cur != __it._M_cur; }
};

///HashTable的vector長度是有講究的,爲了儘量減少衝突(過多的元素被散列
///到同一個桶中),我們的桶個數一般應爲質數個(由於我們是通過hash函數得到
///的值與桶數做mod運算得到需處理元素所在的桶號).這裏取53開始的後28個
///質數,他們中的最大質數大於32位內存可存儲的值.
/// Note: assumes long is at least 32 bits.
enum { __stl_num_primes = 28 };

static const unsigned long __stl_prime_list[__stl_num_primes] =
{
  53ul,         97ul,         193ul,       389ul,       769ul,
  1543ul,       3079ul,       6151ul,      12289ul,     24593ul,
  49157ul,      98317ul,      196613ul,    393241ul,    786433ul,
  1572869ul,    3145739ul,    6291469ul,   12582917ul,  25165843ul,
  50331653ul,   100663319ul,  201326611ul, 402653189ul, 805306457ul,
  1610612741ul, 3221225473ul, 4294967291ul
};

///給定一個值(實際即HashTable中所要存儲的元素個數)得到大於等於
///它的最小質數(實際即HashTable中Array的長度/桶的個數).
inline unsigned long __stl_next_prime(unsigned long __n)
{
  const unsigned long* __first = __stl_prime_list;
  const unsigned long* __last = __stl_prime_list + (int)__stl_num_primes;
  const unsigned long* pos = lower_bound(__first, __last, __n);
  return pos == __last ? *(__last - 1) : *pos;
}

/// Forward declaration of operator==.
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
class hashtable;

template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
bool operator==(const hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>& __ht1,
                const hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>& __ht2);


/// Hashtables handle allocators a bit differently than other containers
///  do.  If we're using standard-conforming allocators, then a hashtable
///  unconditionally has a member variable to hold its allocator, even if
///  it so happens that all instances of the allocator type are identical.
/// This is because, for hashtables, this extra storage is negligible.
///  Additionally, a base class wouldn't serve any other purposes; it
///  wouldn't, for example, simplify the exception-handling code.

template <class _Val, class _Key, class _HashFcn,
          class _ExtractKey, class _EqualKey, class _Alloc>
class hashtable {
public:
  typedef _Key key_type;
  typedef _Val value_type;
  typedef _HashFcn hasher;
  typedef _EqualKey key_equal;

  typedef size_t            size_type;
  typedef ptrdiff_t         difference_type;
  typedef value_type*       pointer;
  typedef const value_type* const_pointer;
  typedef value_type&       reference;
  typedef const value_type& const_reference;

  hasher hash_funct() const { return _M_hash; }
  key_equal key_eq() const { return _M_equals; }

private:
  typedef _Hashtable_node<_Val> _Node;

public:
  typedef _Alloc allocator_type;
  allocator_type get_allocator() const { return allocator_type(); }
private:
  typedef simple_alloc<_Node, _Alloc> _M_node_allocator_type;

  ///分配和回收桶中的一個節點
  _Node* _M_get_node() { return _M_node_allocator_type::allocate(1); }
  void _M_put_node(_Node* __p) { _M_node_allocator_type::deallocate(__p, 1); }

private:
  hasher                _M_hash;
  key_equal             _M_equals;
  _ExtractKey           _M_get_key;
  vector<_Node*,_Alloc> _M_buckets;
  size_type             _M_num_elements;

public:
  typedef _Hashtable_iterator<_Val,_Key,_HashFcn,_ExtractKey,_EqualKey,_Alloc>
          iterator;
  typedef _Hashtable_const_iterator<_Val,_Key,_HashFcn,_ExtractKey,_EqualKey,
                                    _Alloc>
          const_iterator;

  friend struct
  _Hashtable_iterator<_Val,_Key,_HashFcn,_ExtractKey,_EqualKey,_Alloc>;
  friend struct
  _Hashtable_const_iterator<_Val,_Key,_HashFcn,_ExtractKey,_EqualKey,_Alloc>;

public:
  hashtable(size_type __n,
            const _HashFcn&    __hf,
            const _EqualKey&   __eql,
            const _ExtractKey& __ext,
            const allocator_type& __a = allocator_type())
    : _M_hash(__hf),
      _M_equals(__eql),
      _M_get_key(__ext),
      _M_buckets(__a),
      _M_num_elements(0)
  {
    _M_initialize_buckets(__n);
  }

  hashtable(size_type __n,
            const _HashFcn&    __hf,
            const _EqualKey&   __eql,
            const allocator_type& __a = allocator_type())
    :_M_hash(__hf),
      _M_equals(__eql),
      _M_get_key(_ExtractKey()),
      _M_buckets(__a),
      _M_num_elements(0)
  {
    _M_initialize_buckets(__n);
  }

  hashtable(const hashtable& __ht)
    : _M_hash(__ht._M_hash),
      _M_equals(__ht._M_equals),
      _M_get_key(__ht._M_get_key),
      _M_buckets(__ht.get_allocator()),
      _M_num_elements(0)
  {
    _M_copy_from(__ht);
  }

  hashtable& operator= (const hashtable& __ht)
  {
    if (&__ht != this) {
      clear();
      _M_hash = __ht._M_hash;
      _M_equals = __ht._M_equals;
      _M_get_key = __ht._M_get_key;
      _M_copy_from(__ht);
    }
    return *this;
  }

  ~hashtable() { clear(); }

  size_type size() const { return _M_num_elements; }
  size_type max_size() const { return size_type(-1); }
  bool empty() const { return size() == 0; }

  ///STL容器自身的swap成員函數一般都比較高效,HashTable也不例外
  void swap(hashtable& __ht)
  {
    __STD::swap(_M_hash, __ht._M_hash);
    __STD::swap(_M_equals, __ht._M_equals);
    __STD::swap(_M_get_key, __ht._M_get_key);
    _M_buckets.swap(__ht._M_buckets);
    __STD::swap(_M_num_elements, __ht._M_num_elements);
  }

  iterator begin()
  {
      ///找到第一個非空的桶,得到其第一個元素
    for (size_type __n = 0; __n < _M_buckets.size(); ++__n)
      if (_M_buckets[__n])
        return iterator(_M_buckets[__n], this);
    return end();
  }

  iterator end() { return iterator(0, this); }

  const_iterator begin() const
  {
    for (size_type __n = 0; __n < _M_buckets.size(); ++__n)
      if (_M_buckets[__n])
        return const_iterator(_M_buckets[__n], this);
    return end();
  }

  const_iterator end() const { return const_iterator(0, this); }

  friend bool operator==<> (const hashtable&,const hashtable&);

public:

  size_type bucket_count() const { return _M_buckets.size(); }

  size_type max_bucket_count() const
    { return __stl_prime_list[(int)__stl_num_primes - 1]; }

  ///得到第__bucket個桶中的元素個數
  size_type elems_in_bucket(size_type __bucket) const
  {
    size_type __result = 0;
    for (_Node* __cur = _M_buckets[__bucket]; __cur; __cur = __cur->_M_next)
      __result += 1;
    return __result;
  }

  pair<iterator, bool> insert_unique(const value_type& __obj)
  {
    resize(_M_num_elements + 1);
    return insert_unique_noresize(__obj);
  }

  iterator insert_equal(const value_type& __obj)
  {
    resize(_M_num_elements + 1);
    return insert_equal_noresize(__obj);
  }

  pair<iterator, bool> insert_unique_noresize(const value_type& __obj);
  iterator insert_equal_noresize(const value_type& __obj);

  ///下面insert_*中使用到的型別推導技法在STL中很常見
  template <class _InputIterator>
  void insert_unique(_InputIterator __f, _InputIterator __l)
  {
    insert_unique(__f, __l, __ITERATOR_CATEGORY(__f));
  }

  template <class _InputIterator>
  void insert_equal(_InputIterator __f, _InputIterator __l)
  {
    insert_equal(__f, __l, __ITERATOR_CATEGORY(__f));
  }

  template <class _InputIterator>
  void insert_unique(_InputIterator __f, _InputIterator __l,
                     input_iterator_tag)
  {
    for ( ; __f != __l; ++__f)
      insert_unique(*__f);
  }

  template <class _InputIterator>
  void insert_equal(_InputIterator __f, _InputIterator __l,
                    input_iterator_tag)
  {
    for ( ; __f != __l; ++__f)
      insert_equal(*__f);
  }

  template <class _ForwardIterator>
  void insert_unique(_ForwardIterator __f, _ForwardIterator __l,
                     forward_iterator_tag)
  {
    size_type __n = 0;
    distance(__f, __l, __n);
    resize(_M_num_elements + __n);
    for ( ; __n > 0; --__n, ++__f)
      insert_unique_noresize(*__f);
  }

  template <class _ForwardIterator>
  void insert_equal(_ForwardIterator __f, _ForwardIterator __l,
                    forward_iterator_tag)
  {
    size_type __n = 0;
    distance(__f, __l, __n);
    resize(_M_num_elements + __n);
    for ( ; __n > 0; --__n, ++__f)
      insert_equal_noresize(*__f);
  }

  reference find_or_insert(const value_type& __obj);

  iterator find(const key_type& __key)
  {
    size_type __n = _M_bkt_num_key(__key);  ///計算鍵所在的桶序號

    ///在應在的桶中順序查找
    _Node* __first;
    for ( __first = _M_buckets[__n];
          __first && !_M_equals(_M_get_key(__first->_M_val), __key);
          __first = __first->_M_next)
      {}
    return iterator(__first, this);
  }

  const_iterator find(const key_type& __key) const
  {
    size_type __n = _M_bkt_num_key(__key);
    const _Node* __first;
    for ( __first = _M_buckets[__n];
          __first && !_M_equals(_M_get_key(__first->_M_val), __key);
          __first = __first->_M_next)
      {}
    return const_iterator(__first, this);
  }

  size_type count(const key_type& __key) const
  {
    const size_type __n = _M_bkt_num_key(__key);
    size_type __result = 0;

    for (const _Node* __cur = _M_buckets[__n]; __cur; __cur = __cur->_M_next)
      if (_M_equals(_M_get_key(__cur->_M_val), __key))
        ++__result;
    return __result;
  }

  pair<iterator, iterator>
  equal_range(const key_type& __key);

  pair<const_iterator, const_iterator>
  equal_range(const key_type& __key) const;

  size_type erase(const key_type& __key);
  void erase(const iterator& __it);
  void erase(iterator __first, iterator __last);

  void erase(const const_iterator& __it);
  void erase(const_iterator __first, const_iterator __last);

  void resize(size_type __num_elements_hint);
  void clear();

private:
  size_type _M_next_size(size_type __n) const
    { return __stl_next_prime(__n); }

  void _M_initialize_buckets(size_type __n)
  {
    const size_type __n_buckets = _M_next_size(__n); ///計算應分配的桶個數

    _M_buckets.reserve(__n_buckets);    ///分配桶

    ///將每個桶清空
    _M_buckets.insert(_M_buckets.end(), __n_buckets, (_Node*) 0);
    _M_num_elements = 0;
  }

   ///計算鍵所在的桶序號
  size_type _M_bkt_num_key(const key_type& __key) const
  {
    return _M_bkt_num_key(__key, _M_buckets.size());
  }

   ///計算值所在的桶序號
  size_type _M_bkt_num(const value_type& __obj) const
  {
    return _M_bkt_num_key(_M_get_key(__obj));
  }

  size_type _M_bkt_num_key(const key_type& __key, size_t __n) const
  {
      ///通過hash_fun得到的值在和桶個數做mod運算得到
    return _M_hash(__key) % __n;
  }

  size_type _M_bkt_num(const value_type& __obj, size_t __n) const
  {
    return _M_bkt_num_key(_M_get_key(__obj), __n);
  }

  _Node* _M_new_node(const value_type& __obj)
  {
    _Node* __n = _M_get_node();
    __n->_M_next = 0;
   try {
      construct(&__n->_M_val, __obj);
      return __n;
    }catch(...){
        _M_put_node(__n);
    }

  }

  void _M_delete_node(_Node* __n)
  {
    destroy(&__n->_M_val);
    _M_put_node(__n);
  }

  void _M_erase_bucket(const size_type __n, _Node* __first, _Node* __last);
  void _M_erase_bucket(const size_type __n, _Node* __last);

  void _M_copy_from(const hashtable& __ht);

};

template <class _Val, class _Key, class _HF, class _ExK, class _EqK,
          class _All>
_Hashtable_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>&
_Hashtable_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>::operator++()
{
  const _Node* __old = _M_cur;
  _M_cur = _M_cur->_M_next;
  if (!_M_cur) {   ///當前迭代器所指元素爲當前桶中的最後一個元素

    ///得到當前桶序號
    size_type __bucket = _M_ht->_M_bkt_num(__old->_M_val);

    ///從下一個桶開始查找非空桶,查找的的第一個非空桶的第一個元素
    ///即爲所求
    while (!_M_cur && ++__bucket < _M_ht->_M_buckets.size())
      _M_cur = _M_ht->_M_buckets[__bucket];
  }
  return *this;
}

template <class _Val, class _Key, class _HF, class _ExK, class _EqK, class _All>
inline _Hashtable_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>
_Hashtable_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>::operator++(int)
{
  iterator __tmp = *this;
  ++*this;
  return __tmp;
}

template <class _Val, class _Key, class _HF, class _ExK, class _EqK,
          class _All>
_Hashtable_const_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>&
_Hashtable_const_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>::operator++()
{
  const _Node* __old = _M_cur;
  _M_cur = _M_cur->_M_next;
  if (!_M_cur) {
    size_type __bucket = _M_ht->_M_bkt_num(__old->_M_val);
    while (!_M_cur && ++__bucket < _M_ht->_M_buckets.size())
      _M_cur = _M_ht->_M_buckets[__bucket];
  }
  return *this;
}

template <class _Val, class _Key, class _HF, class _ExK, class _EqK,
          class _All>
inline _Hashtable_const_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>
_Hashtable_const_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>::operator++(int)
{
  const_iterator __tmp = *this;
  ++*this;
  return __tmp;
}

///判斷兩個HashTable是否相等
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
bool operator==(const hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>& __ht1,
                const hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>& __ht2)
{
  typedef typename hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::_Node _Node;

  ///(1)桶個數必須相等
  if (__ht1._M_buckets.size() != __ht2._M_buckets.size())
    return false;

  ///(2)每個相同桶序號的相同位置的元素必須相等
  for (int __n = 0; __n < __ht1._M_buckets.size(); ++__n) {
    _Node* __cur1 = __ht1._M_buckets[__n];
    _Node* __cur2 = __ht2._M_buckets[__n];
    for ( ; __cur1 && __cur2 && __cur1->_M_val == __cur2->_M_val;
          __cur1 = __cur1->_M_next, __cur2 = __cur2->_M_next)
      {}
    if (__cur1 || __cur2)
      return false;
  }
  return true;
}

template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
pair<typename hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::iterator, bool>
hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>
  ::insert_unique_noresize(const value_type& __obj)
{
  const size_type __n = _M_bkt_num(__obj);   ///計算應在桶序號
  _Node* __first = _M_buckets[__n];

  ///遍歷該桶,如果已有鍵相同元素,返回
  for (_Node* __cur = __first; __cur; __cur = __cur->_M_next)
    if (_M_equals(_M_get_key(__cur->_M_val), _M_get_key(__obj)))
      return pair<iterator, bool>(iterator(__cur, this), false);

  ///需要插入

  _Node* __tmp = _M_new_node(__obj);   ///構建對應值的插入結點

   ///插入應在桶頭部
  __tmp->_M_next = __first;
  _M_buckets[__n] = __tmp;
  ++_M_num_elements;
  return pair<iterator, bool>(iterator(__tmp, this), true);
}

template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
typename hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::iterator
hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>
  ::insert_equal_noresize(const value_type& __obj)
{
  const size_type __n = _M_bkt_num(__obj);
  _Node* __first = _M_buckets[__n];

  for (_Node* __cur = __first; __cur; __cur = __cur->_M_next)
    if (_M_equals(_M_get_key(__cur->_M_val), _M_get_key(__obj))) {

    ///如果應在桶中已有鍵相同元素,相同元素的最前面
      _Node* __tmp = _M_new_node(__obj);
      __tmp->_M_next = __cur->_M_next;
      __cur->_M_next = __tmp;
      ++_M_num_elements;
      return iterator(__tmp, this);
    }

  ///所在桶中無鍵相同元素
  _Node* __tmp = _M_new_node(__obj);
  __tmp->_M_next = __first;
  _M_buckets[__n] = __tmp;
  ++_M_num_elements;
  return iterator(__tmp, this);
}

template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
typename hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::reference
hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::find_or_insert(const value_type& __obj)
{
    ///首先調整hashtable,爲了防止桶個數太少以至於衝突太多
  resize(_M_num_elements + 1);

  size_type __n = _M_bkt_num(__obj);
  _Node* __first = _M_buckets[__n];

  for (_Node* __cur = __first; __cur; __cur = __cur->_M_next)
    if (_M_equals(_M_get_key(__cur->_M_val), _M_get_key(__obj)))
      return __cur->_M_val;

  _Node* __tmp = _M_new_node(__obj);
  __tmp->_M_next = __first;
  _M_buckets[__n] = __tmp;
  ++_M_num_elements;
  return __tmp->_M_val;
}

template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
pair<typename hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::iterator,
     typename hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::iterator>
hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::equal_range(const key_type& __key)
{
  typedef pair<iterator, iterator> _Pii;
  const size_type __n = _M_bkt_num_key(__key);

  for (_Node* __first = _M_buckets[__n]; __first; __first = __first->_M_next)
    if (_M_equals(_M_get_key(__first->_M_val), __key)) {  ///找到一個鍵與__key相同的元素

      ///繼續從當前位置遍歷該桶,若遇到一個與之不同的鍵即可得到所求
      for (_Node* __cur = __first->_M_next; __cur; __cur = __cur->_M_next)
        if (!_M_equals(_M_get_key(__cur->_M_val), __key))
          return _Pii(iterator(__first, this), iterator(__cur, this));

        ///該桶中鍵與__key相同的元素爲最後一波元素
      for (size_type __m = __n + 1; __m < _M_buckets.size(); ++__m)
        if (_M_buckets[__m])
          return _Pii(iterator(__first, this),iterator(_M_buckets[__m], this));

      return _Pii(iterator(__first, this), end());
    }
  return _Pii(end(), end());
}

template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
pair<typename hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::const_iterator,
     typename hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::const_iterator>
hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>
  ::equal_range(const key_type& __key) const
{
  typedef pair<const_iterator, const_iterator> _Pii;
  const size_type __n = _M_bkt_num_key(__key);

  for (const _Node* __first = _M_buckets[__n] ;
       __first;
       __first = __first->_M_next) {
    if (_M_equals(_M_get_key(__first->_M_val), __key)) {
      for (const _Node* __cur = __first->_M_next;
           __cur;
           __cur = __cur->_M_next)
        if (!_M_equals(_M_get_key(__cur->_M_val), __key))
          return _Pii(const_iterator(__first, this),
                      const_iterator(__cur, this));
      for (size_type __m = __n + 1; __m < _M_buckets.size(); ++__m)
        if (_M_buckets[__m])
          return _Pii(const_iterator(__first, this),
                      const_iterator(_M_buckets[__m], this));
      return _Pii(const_iterator(__first, this), end());
    }
  }
  return _Pii(end(), end());
}

template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
typename hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::size_type
hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::erase(const key_type& __key)
{
 ///找到對應桶以後的刪除操作和單鏈表中刪除元素相同,
 ///最後記着對頭結點需另行處理
  const size_type __n = _M_bkt_num_key(__key);
  _Node* __first = _M_buckets[__n];
  size_type __erased = 0;

  if (__first) {
    _Node* __cur = __first;
    _Node* __next = __cur->_M_next;
    while (__next) {
      if (_M_equals(_M_get_key(__next->_M_val), __key)) {
        __cur->_M_next = __next->_M_next;
        _M_delete_node(__next);
        __next = __cur->_M_next;
        ++__erased;
        --_M_num_elements;
      }
      else {
        __cur = __next;
        __next = __cur->_M_next;
      }
    }

    if (_M_equals(_M_get_key(__first->_M_val), __key)) {
      _M_buckets[__n] = __first->_M_next;
      _M_delete_node(__first);
      ++__erased;
      --_M_num_elements;
    }
  }
  return __erased;
}

template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
void hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::erase(const iterator& __it)
{
    ///同樣找到對應桶後和單鏈表刪除操作相同
  _Node* __p = __it._M_cur;
  if (__p) {
    const size_type __n = _M_bkt_num(__p->_M_val);
    _Node* __cur = _M_buckets[__n];

    if (__cur == __p) {
      _M_buckets[__n] = __cur->_M_next;
      _M_delete_node(__cur);
      --_M_num_elements;
    }
    else {
      _Node* __next = __cur->_M_next;
      while (__next) {
        if (__next == __p) {
          __cur->_M_next = __next->_M_next;
          _M_delete_node(__next);
          --_M_num_elements;
          break;
        }
        else {
          __cur = __next;
          __next = __cur->_M_next;
        }
      }
    }
  }
}

template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
void hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>
  ::erase(iterator __first, iterator __last)
{
  size_type __f_bucket = __first._M_cur ?
    _M_bkt_num(__first._M_cur->_M_val) : _M_buckets.size();

  size_type __l_bucket = __last._M_cur ?
    _M_bkt_num(__last._M_cur->_M_val) : _M_buckets.size();

  if (__first._M_cur == __last._M_cur)
    return;
  else if (__f_bucket == __l_bucket)   ///刪除區間位於同一個桶內
    _M_erase_bucket(__f_bucket, __first._M_cur, __last._M_cur);
  else {
      ///區間內的每個桶分別作合適的刪除
    _M_erase_bucket(__f_bucket, __first._M_cur, 0);
    for (size_type __n = __f_bucket + 1; __n < __l_bucket; ++__n)
      _M_erase_bucket(__n, 0);
    if (__l_bucket != _M_buckets.size())
      _M_erase_bucket(__l_bucket, __last._M_cur);
  }
}

template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
inline void
hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::erase(const_iterator __first,
                                             const_iterator __last)
{
  erase(iterator(const_cast<_Node*>(__first._M_cur),
                 const_cast<hashtable*>(__first._M_ht)),
        iterator(const_cast<_Node*>(__last._M_cur),
                 const_cast<hashtable*>(__last._M_ht)));
}

template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
inline void
hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::erase(const const_iterator& __it)
{
  erase(iterator(const_cast<_Node*>(__it._M_cur),
                 const_cast<hashtable*>(__it._M_ht)));
}

///對hashtable的重新調整,爲了避免桶個數太少以至於衝突太多.這是整個
///hashtable中最關鍵也最複雜的一個成員函數.
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
void hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>
  ::resize(size_type __num_elements_hint)
{
  const size_type __old_n = _M_buckets.size();
  if (__num_elements_hint > __old_n) {
    const size_type __n = _M_next_size(__num_elements_hint);///得到下一個質數
    if (__n > __old_n) {
      vector<_Node*, _All> __tmp(__n, (_Node*)(0),
                                 _M_buckets.get_allocator());   ///重新分配合適大小的vector<list>
      try {
        ///按序號遍歷每個桶
        for (size_type __bucket = 0; __bucket < __old_n; ++__bucket) {

            ///得到桶中第一個元素
          _Node* __first = _M_buckets[__bucket];

          ///依次將該桶中的元素插入到新hashtable中對應的桶中,直到該桶爲空
          while (__first) {
            ///或得該元素在新的hashtable內應在的桶序號
            size_type __new_bucket = _M_bkt_num(__first->_M_val, __n);

            ///將該元素從舊的位置摘下,插入新的hashtable應在的桶內
            _M_buckets[__bucket] = __first->_M_next;
            __first->_M_next = __tmp[__new_bucket];
            __tmp[__new_bucket] = __first;

            ///將first指向舊桶中的第一個元素
            __first = _M_buckets[__bucket];
          }
        }

        ///將得到的新hashtable和原有hashtable替換
        _M_buckets.swap(__tmp);
      }
      catch(...) {
          ///如果操作失敗,需要依次刪除所有新hashtable內的元素,
          ///以防內存泄露
        for (size_type __bucket = 0; __bucket < __tmp.size(); ++__bucket) {
          while (__tmp[__bucket]) {
            _Node* __next = __tmp[__bucket]->_M_next;
            _M_delete_node(__tmp[__bucket]);
            __tmp[__bucket] = __next;
          }
        }
        throw;
      }
    }
  }
}

///刪除序號爲__n的桶內[first,last)區間內的元素
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
void hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>
  ::_M_erase_bucket(const size_type __n, _Node* __first, _Node* __last)
{
  _Node* __cur = _M_buckets[__n];

  ///從頭結點開始刪除,需要特殊處理
  if (__cur == __first)
    _M_erase_bucket(__n, __last);
  else {
    _Node* __next;

    ///找到需要刪除的起點
    for (__next = __cur->_M_next;
         __next != __first;
         __cur = __next, __next = __cur->_M_next)
      ;

    ///類似單鏈表刪除
    while (__next != __last) {
      __cur->_M_next = __next->_M_next;
      _M_delete_node(__next);
      __next = __cur->_M_next;
      --_M_num_elements;
    }
  }
}

///刪除序號爲__n的桶內自頭結點至__last的元素,不包括__last
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
void hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>
  ::_M_erase_bucket(const size_type __n, _Node* __last)
{
  _Node* __cur = _M_buckets[__n];
  while (__cur != __last) {
    _Node* __next = __cur->_M_next;
    _M_delete_node(__cur);
    __cur = __next;
    _M_buckets[__n] = __cur;   ///記着調整_M_buckets[__n]的指向
    --_M_num_elements;
  }
}

template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
void hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::clear()
{
    ///挨個清空單鏈表,並將桶清空
  for (size_type __i = 0; __i < _M_buckets.size(); ++__i) {
    _Node* __cur = _M_buckets[__i];
    while (__cur != 0) {
      _Node* __next = __cur->_M_next;
      _M_delete_node(__cur);
      __cur = __next;
    }
    _M_buckets[__i] = 0;
  }
  _M_num_elements = 0;
}


template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
void hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>
  ::_M_copy_from(const hashtable& __ht)
{
    ///挨個複製單鏈表
  _M_buckets.clear();
  _M_buckets.reserve(__ht._M_buckets.size());
  _M_buckets.insert(_M_buckets.end(), __ht._M_buckets.size(), (_Node*) 0);
  try {
    for (size_type __i = 0; __i < __ht._M_buckets.size(); ++__i) {
      const _Node* __cur = __ht._M_buckets[__i];
      if (__cur) {
        _Node* __copy = _M_new_node(__cur->_M_val);
        _M_buckets[__i] = __copy;

        for (_Node* __next = __cur->_M_next;
             __next;
             __cur = __next, __next = __cur->_M_next) {
          __copy->_M_next = _M_new_node(__next->_M_val);
          __copy = __copy->_M_next;
        }
      }
    }
    _M_num_elements = __ht._M_num_elements;
  }catch(...){
      clear();
      throw;
  }
}

hash_table是STL中hash_map 和 hash_set 的內部數據結構,hash_table的插入/刪除/查找的時間複雜度都爲O(1),是查找速度最快的一種數據結構,但是hash_table中的數據是無序的,一般也只有在數據不需要排序,只需要滿足快速查找/插入/刪除的時候使用hash_table。hash_table的擴展是將原hash_table中的數據摘下來插入到一個臨時的hash_table中,因爲每個桶都使用list來實現的,因此插入刪除都不存在內存copy,所以也是很高效的,最後再將臨時hash_table和原來的hash_table(此時已經爲空)交換。

發表評論
所有評論
還沒有人評論,想成為第一個評論的人麼? 請在上方評論欄輸入並且點擊發布.
相關文章