从零开始写STL—哈希表
2024-09-08 12:37:14
static const int _stl_num_primes = 28;
template<typename T, typename Hash = xhash<T>>
class hashtable;
ministl::vector<int> prime_list =
{
53, 97, 193, 389, 769, 1543, 3079, 6151, 12289, 24593, 49157, 98317, 196613, 393241,
786433, 1572869, 3145739, 6291469, 12582917, 25165843, 50331653, 100663319, 201326611,
402653189, 805306457, 1610612741, 3221225473, 4294967291,
};
template<typename T>
class hash_iterator : public bidirectional_iterator<T>
{
public:
hashtable<T>* hash_ptr;
list_iterator<T> node_ptr;
list<T>* bucket_ptr;
hash_iterator(hashtable<T>* hp, list<T>* bp, list_iterator<T> np) :hash_ptr(hp), bucket_ptr(bp), node_ptr(np) {}
hash_iterator& operator++()
{
if (++node_ptr == bucket_ptr->end())
{
auto tmp = node_ptr;
--tmp;
size_t cur = hash_ptr->hash(*tmp, hash_ptr->bucket_size()) + 1;
while (cur < hash_ptr->bucket_size() && (*hash_ptr).data[cur].empty())
cur++;
bucket_ptr = cur < hash_ptr->bucket_size() ? &(hash_ptr->data[cur]) : hash_ptr->data.end();
node_ptr = (bucket_ptr == hash_ptr->data.end() ? nullptr : bucket_ptr->begin());
}
return *this;
}
hash_iterator operator++(int)
{
hash_iterator tmp = *this;
++*this;
return tmp;
}
hash_iterator& operator--()
{
if (node_ptr-- = bucket_ptr->begin())
{
size_t cur = hash_ptr->hash(*node_ptr, hash_ptr->bucket_size()) - 1;
if (cur < 0) cur = 0;
else
{
while (cur >= 0 && (*hash_ptr).data[cur].empty())
cur--;
}
bucket_ptr = &(hash_ptr->data[cur]);
node_ptr = bucket_ptr->begin();
}
return *this;
}
hash_iterator operator--(int)
{
hash_iterator tmp = *this;
--*this;
return tmp;
}
T& operator* ()
{
return *node_ptr;
}
bool operator==(const hash_iterator& rhs)
{
return hash_ptr == rhs.hash_ptr && node_ptr == rhs.node_ptr && bucket_ptr == rhs.bucket_ptr;
}
bool operator!=(const hash_iterator& rhs)
{
return !(*this == rhs);
}
};
template<typename T,typename Hash = xhash<T> >
class hashtable
{
friend class hash_iterator<T>;
public:
typedef T value_type;
typedef Hash hash_type;
typedef vector<list<T>> table;
typedef hash_iterator<T> iterator;
private:
Hash hash;
table data;
size_t elem_num;//the sum of all elements
pair<iterator, bool> _insert_unique_noresize(T val)
{
size_t index = hash(val, data.size());
for (auto it = data[index].begin(); it != data[index].end(); it++)
{
if (*it == val)
return make_pair<iterator, bool >(iterator(this, &data[index], it), false);
}
data[index].push_front(val);
elem_num++;
return make_pair<iterator, bool>(iterator(this, &data[index], data[index].begin()), true);
}
iterator _insert_equal_noresize(T val)
{
size_t index = hash(val, data.size());
for (auto it = data[index].begin(); it != data[index].end(); it++)
{
if (*it == val)
{
return iterator(this, &data[index], data[index].insert(it, val));
}
}
data[index].push_front(val);
elem_num++;
return iterator(this, &data[index], data[index].begin());
}
void initialize_buckets(size_t n)
{
data.resize(n);
elem_num = 0;
}
public:
hashtable()
{
initialize_buckets(53);
}
hashtable(size_t n)
{
initialize_buckets(next_prime(n));
}
void clear()
{
for (auto& it : data)
{
it.clear();
}
elem_num = 0;
}
bool empty()
{
return elem_num == 0;
}
pair<iterator, bool>find(const T& val)
{
size_t index = hash(val, data.size());
for (auto it = data[index].begin(); it != data[index].end(); it++)
{
if (*it == val)
return ministl::make_pair<iterator, bool >(iterator(this, &data[index], it), true);
}
return ministl::make_pair<iterator, bool >(end(), false);
}
iterator begin() const
{
auto it = data.begin();
while (it != data.end() && it->empty())
it++;
return iterator(const_cast<hashtable<T>*>(this), it, it == data.end() ? nullptr : it->begin());
}
iterator begin()
{
auto it = data.begin();
while (it != data.end() && it->empty())
it++;
return iterator(this, it, it == data.end() ? nullptr : it->begin());
}
iterator end() const
{
return iterator(const_cast<hashtable<T>*>(this), data.end(), nullptr);
}
iterator end()
{
return iterator(this, data.end(), nullptr);
}
inline unsigned long next_prime(unsigned long n)
{
for (auto it : prime_list)
{
if (it >= n)
return n;
}
return prime_list[_stl_num_primes - 1];
}
size_t bucket_size()
{
return data.size();
}
size_t bucket_size() const
{
return data.size();
}
size_t size()
{
return elem_num;
}
size_t hash_val(const T& val)
{
return hash(val, data.size());
}
pair<iterator, bool> insert_unique(T val)
{
rehash(elem_num + 1);
return _insert_unique_noresize(val);
}
iterator insert_equal(T val)
{
rehash(elem_num + 1);
return _insert_equal_noresize(val);
}
iterator erase_iter(iterator pos)
{
iterator ret = pos;
ret++;
pos.bucket_ptr->erase(pos.node_ptr);
return ret;
}
void erase_val(const T& val)
{
size_t index = hash(val, data.size());
for (auto it = data[index].begin(); it != data[index].end();)
{
if (*it == val)
it = data[index].erase(it);
else
it++;
}
}
void rehash(size_t n)
{
if (n > bucket_size())
{
elem_num = elem_num ? elem_num:1;
while (n > elem_num) elem_num *= 2;
size_t new_size = next_prime(elem_num);
hashtable new_table(new_size);
if (!empty())
{
iterator it = begin();
for (; it != end(); it++)
{
//std::cout << *it << std::endl;
new_table.insert_equal(*it);
}
}
swap(*this, new_table);
}
}
hashtable& operator=(const hashtable& rhs)
{
clear();
for (auto it = rhs.begin(); it != rhs.end(); it++)
{
insert_equal(*it);
}
return *this;
}
};
template<typename T>
void swap(const hashtable<T>& lhs, const hashtable<T>& rhs)
{
using ministl::swap;
swap(lhs.data, rhs.data);
swap(lhs.elem_num, rhs.elem_num);
swap(lhs.hash, rhs.hash);
}
template<typename T, typename Hash = xhash<T> >
class unordered_set
{
public:
typedef T value_type;
typedef Hash hash_type;
typedef vector<list<T>> table;
typedef hash_iterator<T> iterator;
typedef hashtable<T> seq;
private:
seq con;
public:
unordered_set() :con() {}
unordered_set(std::initializer_list<T> l)
{
unordered_set();
insert(l.begin(), l.end());
}
iterator begin()
{
return con.begin();
}
iterator end()
{
return con.end();
}
iterator find(const value_type& key)
{
return con.find(key).first;
}
iterator insert(const value_type& key)
{
return con.insert_unique(key).first;
}
iterator insert(value_type& key)
{
return con.insert_unique(key);
}
template <class InputIterator>
void insert(InputIterator first, InputIterator last)
{
for (auto it = first; it != last; it++)
insert(*it);
}
size_t erase(const value_type& key)
{
size_t num = count(key);
if(num)
con.erase_val(key);
return num;
}
iterator erase(const iterator it)
{
return con.erase_iter(it);
}
iterator erase(const iterator first, const iterator last)
{
iterator it = first;
for (; it != last;)
it = con.erase_iter(it);
return it;
}
size_t count(const value_type& val)
{
return (con.find(val).second == true) ? 1 : 0;
}
bool empty()
{
return con.empty();
}
size_t size()
{
return con.size();
}
void clear()
{
con.clear();
}
void swap(unordered_set<T>& rhs)
{
using ministl::swap;
swap(con, rhs.con);
}
size_t bucket_count()
{
return con.bucket_size();
}
void rehash()
{
con.rehash(con.elem_num + 1);
}
double load_factor()
{
return 1.0 * con.elem_num / bucket_count();
}
};
//Todo: map_pair的hashval
//模板特化无法匹配？
template<typename K , typename T, typename Hash = xhash<map_pair<K,T>> >
class unordered_map
{
public:
typedef K key_type;
typedef T value_type;
typedef Hash hash_type;
typedef vector<list<map_pair<K,T>>> table;
typedef hash_iterator<map_pair<K,T>> iterator;
typedef hashtable<map_pair<K, T>, xhash<map_pair<K, T>> > seq;
private:
seq con;
public:
unordered_map() :con() {}
iterator begin()
{
return con.begin();
}
iterator begin() const
{
return con.begin();
}
iterator end()
{
return con.end();
}
iterator end() const
{
return con.end();
}
value_type& operator[](const key_type& key)
{
map_pair<K, T> p(key, T());
auto tmp = con.find(p);
iterator it = tmp.first;
if (tmp.second == false)
{
it = con.insert_equal(p);
}
return (*it).second;
}
value_type& at(const key_type& key)
{
auto tmp = con.find(key);
if (tmp.second == false)
{
std::cerr << "unordered_map out of range \n";
std::exit(1);
}
return (*it).second;
}
const value_type& at(const key_type& key) const
{
auto tmp = con.find(key);
if (tmp.second == false)
{
std::cerr << "unordered_map out of range \n";
std::exit(1);
}
return (*it).second;
}
iterator find(const key_type& key)
{
return con.find(key).first;
}
pair<iterator, iterator>
equal_range(const key_type& k)
{
auto first = con.find(k).first;
if (first == con.end())
return make_pair<iterator, iterator>(first, first);
else
{
iterator last = ++first;
while (last != con.end() && (*last == *first))
{
last++;
}
return make_pair<iterator, iterator>(first, last);
}
}
size_t count(const key_type& k)
{
auto first = con.find(k).first;
if (first == con.end())
return 0;
else
{
iterator last = ++first;
size_t cnt = 1;
while (last != con.end() && (*last == *first))
{
last++, cnt++;
}
return cnt;
}
}
auto insert(const map_pair<K, T> p)
{
return con.insert_unique(p);
}
size_t erase(const value_type& key)
{
size_t num = count(key);
if (num)
con.erase_val(key);
return num;
}
iterator erase(const iterator it)
{
return con.erase_iter(it);
}
iterator erase(const iterator first, const iterator last)
{
iterator it = first;
for (; it != last;)
it = con.erase_iter(it);
return it;
}
void rehash(size_t n)
{
con.rehash(n);
}
};
最新文章
- MVC MODEL Attribute 操纵速记
- 微信公众平台开发（71）OAuth2.0网页授权
- Linux学习笔记（7）Linux常用命令之压缩解压命令
- 多列布局——column-width
- POJ 2699 The Maximum Number of Strong Kings Description
- 如何使用编辑模板在ASPxGridView中进行新增修改（除去常规的gridviw模板编辑外）
- 一份Java学习路线图
- APP-SQLAP-10771
- MongoDB时间处理问题
- DTCoreText
- c++STL(栈、队列）
- Docker（十二）：Docker集群管理之Compose
- 关于java字节流的read()方法返回值为int的思考
- HDU 3565 Bi-peak Number（数位DP）题解
- puppeteer（二）操作实例——新Web自动化工具更轻巧更简单
- 设置DataGridView中表头颜色
- loj #117. 有源汇有上下界最小流
- Centos7安装Openvpn
- JVM相关笔记
- POJ 1458 Common Subsequence(最长公共子序列LCS)