用二次探测实现

时间:2013-04-24 05:37:52

标签: c++ hash

我有这个代码,我想用二次探测方法重新实现它,我的算法是 i = (i + count) % CAPACITY;。我不确定我应该怎么做,所以帮助会很好。我想你只是改变哈希函数和next_index函数,我不太确定。这是我需要重新实现的代码。底部顶部模板上的头文件。

#ifndef TABLE1_H
#define TABLE1_H
#include <cstdlib>    // Provides size_t

namespace main_savitch_12A
{
template <class RecordType>
class table
{
public:
    // MEMBER CONSTANT -- See Appendix E if this fails to compile.
    static const std::size_t CAPACITY = 811;
    // CONSTRUCTOR
    table( );
    // MODIFICATION MEMBER FUNCTIONS
    void insert(const RecordType& entry);
    void remove(int key);
    // CONSTANT MEMBER FUNCTIONS
    bool is_present(int key) const;
    void find(int key, bool& found, RecordType& result) const;
    std::size_t size( ) const { return used; }
private:
    // MEMBER CONSTANTS -- These are used in the key field of special records.
    static const int NEVER_USED = -1;
    static const int PREVIOUSLY_USED = -2;
    // MEMBER VARIABLES
    RecordType data[CAPACITY];
    std::size_t used;
    // HELPER FUNCTIONS
    std::size_t hash(int key) const;
    std::size_t next_index(std::size_t index) const;
    void find_index(int key, bool& found, std::size_t& index) const;
    bool never_used(std::size_t index) const;
    bool is_vacant(std::size_t index) const;
};
}
 #include "table1.template" // Include the implementation.
 #endif
 //End Of Header




#include <cassert>  // Provides assert
#include <cstdlib>  // Provides size_t

namespace main_savitch_12A
{
template <class RecordType>
const std::size_t table<RecordType>::CAPACITY; 

template <class RecordType>
const int table<RecordType>::NEVER_USED;

template <class RecordType>
const int table<RecordType>::PREVIOUSLY_USED;

template <class RecordType>
table<RecordType>::table( )
{
    std::size_t i;

    used = 0;
    for (i = 0; i < CAPACITY; ++i)
        data[i].key = NEVER_USED;  // Indicates a spot that's never been used.
}

template <class RecordType>
void table<RecordType>::insert(const RecordType& entry)
// Library facilities used: cassert
{
    bool already_present;   // True if entry.key is already in the table
    std::size_t index;        // data[index] is location for the new entry

    assert(entry.key >= 0);

    // Set index so that data[index] is the spot to place the new entry.
    find_index(entry.key, already_present, index);

    // If the key wasn't already there, then find the location for the new entry.
    if (!already_present)
    {
        assert(size( ) < CAPACITY);
        index = hash(entry.key);
        while (!is_vacant(index))
            index = next_index(index);
        ++used;
    }

    data[index] = entry;
}

template <class RecordType>
void table<RecordType>::remove(int key)
// Library facilities used: cassert
{
    bool found;        // True if key occurs somewhere in the table
    std::size_t index;   // Spot where data[index].key == key

    assert(key >= 0);

    find_index(key, found, index);
    if (found)
    {   // The key was found, so remove this record and reduce used by 1.
        data[index].key = PREVIOUSLY_USED; // Indicates a spot that's no longer in use.
        --used;
    }
}

template <class RecordType>
bool table<RecordType>::is_present(int key) const
// Library facilities used: assert.h
{
    bool found;
    std::size_t index;

    assert(key >= 0);

    find_index(key, found, index);
    return found;
}

template <class RecordType>
void table<RecordType>::find(int key, bool& found, RecordType& result) const
// Library facilities used: cassert.h
{
    std::size_t index;

    assert(key >= 0);

    find_index(key, found, index);
    if (found)
        result = data[index];
}

template <class RecordType>
inline std::size_t table<RecordType>::hash(int key) const
{
    return (key % CAPACITY);
}

template <class RecordType>
inline std::size_t table<RecordType>::next_index(std::size_t index) const
// Library facilities used: cstdlib
{
    return ((index+1) % CAPACITY);
}

template <class RecordType>
void table<RecordType>::find_index(int key, bool& found, std::size_t& i) const
// Library facilities used: cstdlib
{
std::size_t count; // Number of entries that have been examined

count = 0;
i = hash(key);
while((count < CAPACITY) && (data[i].key != NEVER_USED) && (data[i].key != key))
{
    ++count;
    i = next_index(i);
}
found = (data[i].key == key);
}

template <class RecordType>
inline bool table<RecordType>::never_used(std::size_t index) const
{
return (data[index].key == NEVER_USED);
}

template <class RecordType>
inline bool table<RecordType>::is_vacant(std::size_t index) const
{
return (data[index].key == NEVER_USED) || (data[index].key == PREVIOUSLY_USED);
}
}

1 个答案:

答案 0 :(得分:2)

使用此代码,您正在进行线性探测

    index = hash(entry.key);
    while (!is_vacant(index))
        index = next_index(index);

template <class RecordType>
inline std::size_t table<RecordType>::next_index(std::size_t index) const
// Library facilities used: cstdlib
{
    return ((index+1) % CAPACITY);
}

假设您的地图几乎已满,哈希返回23,那么您要测试的下一个插槽将分别为24,25,26,27等。

关于二次探测的所有不同之处是当插槽已满时要测试的插槽模式。再次假设哈希返回23,那么下一个要测试的槽将是23 + 1 = 24,下一个将是23 + 4 = 27,下一个将是23 + 9 = 32,下一个将是23 + 16 = 39。看模式?每次测试23 + n * n。这是二次探测。当然,所有值都应该是mod CAPACITY,就像你现在所做的那样。

换句话说,您不需要更改散列函数,只需更改插入内的while循环。