mirror of
https://github.com/krahets/hello-algo.git
synced 2024-12-26 11:46:29 +08:00
513 lines
13 KiB
Markdown
513 lines
13 KiB
Markdown
---
|
||
comments: true
|
||
---
|
||
|
||
# 哈希表
|
||
|
||
哈希表通过建立「键 key」和「值 value」之间的映射,实现高效的元素查找。具体地,输入一个 key ,在哈希表中查询并获取 value ,时间复杂度为 $O(1)$ 。
|
||
|
||
例如,给定一个包含 $n$ 个学生的数据库,每个学生有“姓名 `name` ”和“学号 `id` ”两项数据,希望实现一个查询功能:**输入一个学号,返回对应的姓名**,则可以使用哈希表实现。
|
||
|
||
![hash_map](hash_map.assets/hash_map.png)
|
||
|
||
<p align="center"> Fig. 哈希表抽象表示 </p>
|
||
|
||
## 哈希表优势
|
||
|
||
除了哈希表之外,还可以使用以下数据结构来实现上述查询功能:
|
||
|
||
1. **无序数组:** 每个元素为 `[学号, 姓名]` ;
|
||
2. **有序数组:** 将 `1.` 中的数组按照学号从小到大排序;
|
||
3. **链表:** 每个结点的值为 `[学号, 姓名]` ;
|
||
4. **二叉搜索树:** 每个结点的值为 `[学号, 姓名]` ,根据学号大小来构建树;
|
||
|
||
使用上述方法,各项操作的时间复杂度如下表所示(在此不做赘述,详解可见 [二叉搜索树章节](https://www.hello-algo.com/chapter_tree/binary_search_tree/#_6))。无论是查找元素、还是增删元素,哈希表的时间复杂度都是 $O(1)$ ,全面胜出!
|
||
|
||
<div class="center-table" markdown>
|
||
|
||
| | 无序数组 | 有序数组 | 链表 | 二叉搜索树 | 哈希表 |
|
||
| -------- | -------- | ----------- | ------ | ----------- | ------ |
|
||
| 查找元素 | $O(n)$ | $O(\log n)$ | $O(n)$ | $O(\log n)$ | $O(1)$ |
|
||
| 插入元素 | $O(1)$ | $O(n)$ | $O(1)$ | $O(\log n)$ | $O(1)$ |
|
||
| 删除元素 | $O(n)$ | $O(n)$ | $O(n)$ | $O(\log n)$ | $O(1)$ |
|
||
|
||
</div>
|
||
|
||
## 哈希表常用操作
|
||
|
||
哈希表的基本操作包括 **初始化、查询操作、添加与删除键值对**。
|
||
|
||
=== "Java"
|
||
|
||
```java title="hash_map.java"
|
||
/* 初始化哈希表 */
|
||
Map<Integer, String> map = new HashMap<>();
|
||
|
||
/* 添加操作 */
|
||
// 在哈希表中添加键值对 (key, value)
|
||
map.put(12836, "小哈");
|
||
map.put(15937, "小啰");
|
||
map.put(16750, "小算");
|
||
map.put(13276, "小法");
|
||
map.put(10583, "小鸭");
|
||
|
||
/* 查询操作 */
|
||
// 向哈希表输入键 key ,得到值 value
|
||
String name = map.get(15937);
|
||
|
||
/* 删除操作 */
|
||
// 在哈希表中删除键值对 (key, value)
|
||
map.remove(10583);
|
||
```
|
||
|
||
=== "C++"
|
||
|
||
```cpp title="hash_map.cpp"
|
||
/* 初始化哈希表 */
|
||
unordered_map<int, string> map;
|
||
|
||
/* 添加操作 */
|
||
// 在哈希表中添加键值对 (key, value)
|
||
map[12836] = "小哈";
|
||
map[15937] = "小啰";
|
||
map[16750] = "小算";
|
||
map[13276] = "小法";
|
||
map[10583] = "小鸭";
|
||
|
||
/* 查询操作 */
|
||
// 向哈希表输入键 key ,得到值 value
|
||
string name = map[15937];
|
||
|
||
/* 删除操作 */
|
||
// 在哈希表中删除键值对 (key, value)
|
||
map.erase(10583);
|
||
```
|
||
|
||
=== "Python"
|
||
|
||
```python title="hash_map.py"
|
||
""" 初始化哈希表 """
|
||
mapp = {}
|
||
|
||
""" 添加操作 """
|
||
# 在哈希表中添加键值对 (key, value)
|
||
mapp[12836] = "小哈"
|
||
mapp[15937] = "小啰"
|
||
mapp[16750] = "小算"
|
||
mapp[13276] = "小法"
|
||
mapp[10583] = "小鸭"
|
||
|
||
""" 查询操作 """
|
||
# 向哈希表输入键 key ,得到值 value
|
||
name = mapp[15937]
|
||
|
||
""" 删除操作 """
|
||
# 在哈希表中删除键值对 (key, value)
|
||
mapp.pop(10583)
|
||
```
|
||
|
||
=== "Go"
|
||
|
||
```go title="hash_map_test.go"
|
||
/* 初始化哈希表 */
|
||
mapp := make(map[int]string)
|
||
|
||
/* 添加操作 */
|
||
// 在哈希表中添加键值对 (key, value)
|
||
mapp[12836] = "小哈"
|
||
mapp[15937] = "小啰"
|
||
mapp[16750] = "小算"
|
||
mapp[13276] = "小法"
|
||
mapp[10583] = "小鸭"
|
||
|
||
/* 查询操作 */
|
||
// 向哈希表输入键 key ,得到值 value
|
||
name := mapp[15937]
|
||
|
||
/* 删除操作 */
|
||
// 在哈希表中删除键值对 (key, value)
|
||
delete(mapp, 10583)
|
||
```
|
||
|
||
=== "JavaScript"
|
||
|
||
```js title="hash_map.js"
|
||
|
||
```
|
||
|
||
=== "TypeScript"
|
||
|
||
```typescript title="hash_map.ts"
|
||
|
||
```
|
||
|
||
=== "C"
|
||
|
||
```c title="hash_map.c"
|
||
|
||
```
|
||
|
||
=== "C#"
|
||
|
||
```csharp title="hash_map.cs"
|
||
|
||
```
|
||
|
||
遍历哈希表有三种方式,即 **遍历键值对、遍历键、遍历值**。
|
||
|
||
=== "Java"
|
||
|
||
```java title="hash_map.java"
|
||
/* 遍历哈希表 */
|
||
// 遍历键值对 key->value
|
||
for (Map.Entry <Integer, String> kv: map.entrySet()) {
|
||
System.out.println(kv.getKey() + " -> " + kv.getValue());
|
||
}
|
||
// 单独遍历键 key
|
||
for (int key: map.keySet()) {
|
||
System.out.println(key);
|
||
}
|
||
// 单独遍历值 value
|
||
for (String val: map.values()) {
|
||
System.out.println(val);
|
||
}
|
||
```
|
||
|
||
=== "C++"
|
||
|
||
```cpp title="hash_map.cpp"
|
||
/* 遍历哈希表 */
|
||
// 遍历键值对 key->value
|
||
for (auto kv: map) {
|
||
cout << kv.first << " -> " << kv.second << endl;
|
||
}
|
||
// 单独遍历键 key
|
||
for (auto key: map) {
|
||
cout << key.first << endl;
|
||
}
|
||
// 单独遍历值 value
|
||
for (auto val: map) {
|
||
cout << val.second << endl;
|
||
}
|
||
```
|
||
|
||
=== "Python"
|
||
|
||
```python title="hash_map.py"
|
||
""" 遍历哈希表 """
|
||
# 遍历键值对 key->value
|
||
for key, value in mapp.items():
|
||
print(key, "->", value)
|
||
# 单独遍历键 key
|
||
for key in mapp.keys():
|
||
print(key)
|
||
# 单独遍历值 value
|
||
for value in mapp.values():
|
||
print(value)
|
||
```
|
||
|
||
=== "Go"
|
||
|
||
```go title="hash_map_test.go"
|
||
/* 遍历哈希表 */
|
||
// 遍历键值对 key->value
|
||
for key, value := range mapp {
|
||
fmt.Println(key, "->", value)
|
||
}
|
||
// 单独遍历键 key
|
||
for key := range mapp {
|
||
fmt.Println(key)
|
||
}
|
||
// 单独遍历值 value
|
||
for _, value := range mapp {
|
||
fmt.Println(value)
|
||
}
|
||
```
|
||
|
||
=== "JavaScript"
|
||
|
||
```js title="hash_map.js"
|
||
|
||
```
|
||
|
||
=== "TypeScript"
|
||
|
||
```typescript title="hash_map.ts"
|
||
|
||
```
|
||
|
||
=== "C"
|
||
|
||
```c title="hash_map.c"
|
||
|
||
```
|
||
|
||
=== "C#"
|
||
|
||
```csharp title="hash_map.cs"
|
||
|
||
```
|
||
|
||
## 哈希函数
|
||
|
||
哈希表中存储元素的数据结构被称为「桶 Bucket」,底层实现可能是数组、链表、二叉树(红黑树),或是它们的组合。
|
||
|
||
最简单地,**我们可以仅用一个「数组」来实现哈希表**。首先,将所有 value 放入数组中,那么每个 value 在数组中都有唯一的「索引」。显然,访问 value 需要给定索引,而为了 **建立 key 和索引之间的映射关系**,我们需要使用「哈希函数 Hash Function」。
|
||
|
||
设数组为 `bucket` ,哈希函数为 `f(x)` ,输入键为 `key` 。那么获取 value 的步骤为:
|
||
|
||
1. 通过哈希函数计算出索引,即 `index = f(key)` ;
|
||
2. 通过索引在数组中获取值,即 `value = bucket[index]` ;
|
||
|
||
以上述学生数据 `key 学号 -> value 姓名` 为例,我们可以将「哈希函数」设计为
|
||
|
||
$$
|
||
f(x) = x \% 100
|
||
$$
|
||
|
||
![hash_function](hash_map.assets/hash_function.png)
|
||
|
||
<p align="center"> Fig. 哈希函数 </p>
|
||
|
||
=== "Java"
|
||
|
||
```java title="array_hash_map.java"
|
||
/* 键值对 int->String */
|
||
class Entry {
|
||
public int key; // 键
|
||
public String val; // 值
|
||
public Entry(int key, String val) {
|
||
this.key = key;
|
||
this.val = val;
|
||
}
|
||
}
|
||
|
||
/* 基于数组简易实现的哈希表 */
|
||
class ArrayHashMap {
|
||
private List<Entry> bucket;
|
||
public ArrayHashMap() {
|
||
// 初始化一个长度为 100 的桶(数组)
|
||
bucket = new ArrayList<>();
|
||
for (int i = 0; i < 100; i++) {
|
||
bucket.add(null);
|
||
}
|
||
}
|
||
|
||
/* 哈希函数 */
|
||
private int hashFunc(int key) {
|
||
int index = key % 100;
|
||
return index;
|
||
}
|
||
|
||
/* 查询操作 */
|
||
public String get(int key) {
|
||
int index = hashFunc(key);
|
||
Entry pair = bucket.get(index);
|
||
if (pair == null) return null;
|
||
return pair.val;
|
||
}
|
||
|
||
/* 添加操作 */
|
||
public void put(int key, String val) {
|
||
Entry pair = new Entry(key, val);
|
||
int index = hashFunc(key);
|
||
bucket.set(index, pair);
|
||
}
|
||
|
||
/* 删除操作 */
|
||
public void remove(int key) {
|
||
int index = hashFunc(key);
|
||
// 置为 null,代表删除
|
||
bucket.set(index, null);
|
||
}
|
||
}
|
||
```
|
||
|
||
=== "C++"
|
||
|
||
```cpp title="array_hash_map.cpp"
|
||
/* 键值对 int->String */
|
||
struct Entry {
|
||
public:
|
||
int key;
|
||
string val;
|
||
Entry(int key, string val) {
|
||
this->key = key;
|
||
this->val = val;
|
||
}
|
||
};
|
||
|
||
/* 基于数组简易实现的哈希表 */
|
||
class ArrayHashMap {
|
||
private:
|
||
vector<Entry*> bucket;
|
||
public:
|
||
ArrayHashMap() {
|
||
// 初始化一个长度为 100 的桶(数组)
|
||
bucket= vector<Entry*>(100);
|
||
}
|
||
|
||
/* 哈希函数 */
|
||
int hashFunc(int key) {
|
||
int index = key % 100;
|
||
return index;
|
||
}
|
||
|
||
/* 查询操作 */
|
||
string get(int key) {
|
||
int index = hashFunc(key);
|
||
Entry* pair = bucket[index];
|
||
return pair->val;
|
||
}
|
||
|
||
/* 添加操作 */
|
||
void put(int key, string val) {
|
||
Entry* pair = new Entry(key, val);
|
||
int index = hashFunc(key);
|
||
bucket[index] = pair;
|
||
}
|
||
|
||
/* 删除操作 */
|
||
void remove(int key) {
|
||
int index = hashFunc(key);
|
||
// 置为空字符,代表删除
|
||
bucket[index] = nullptr;
|
||
}
|
||
};
|
||
```
|
||
|
||
=== "Python"
|
||
|
||
```python title="array_hash_map.py"
|
||
""" 键值对 int->String """
|
||
class Entry:
|
||
def __init__(self, key, val):
|
||
self.key = key
|
||
self.val = val
|
||
|
||
""" 基于数组简易实现的哈希表 """
|
||
class ArrayHashMap:
|
||
def __init__(self):
|
||
# 初始化一个长度为 100 的桶(数组)
|
||
self.bucket = [None] * 100
|
||
|
||
""" 哈希函数 """
|
||
def hashFunc(self, key):
|
||
index = key % 100
|
||
return index
|
||
|
||
""" 查询操作 """
|
||
def get(self, key):
|
||
index = self.hashFunc(key)
|
||
pair = self.bucket[index]
|
||
if pair is None:
|
||
return None
|
||
return pair.val
|
||
|
||
""" 添加操作 """
|
||
def put(self, key, val):
|
||
pair = Entry(key, val)
|
||
index = self.hashFunc(key)
|
||
self.bucket[index] = pair
|
||
|
||
""" 删除操作 """
|
||
def remove(self, key):
|
||
index = self.hashFunc(key)
|
||
# 置为空字符,代表删除
|
||
self.bucket[index] = None
|
||
```
|
||
|
||
=== "Go"
|
||
|
||
```go title="array_hash_map.go"
|
||
/* 键值对 int->String */
|
||
type Entry struct {
|
||
key int
|
||
val string
|
||
}
|
||
|
||
/* 基于数组简易实现的哈希表 */
|
||
type ArrayHashMap struct {
|
||
bucket []*Entry
|
||
}
|
||
|
||
func newArrayHashMap() *ArrayHashMap {
|
||
// 初始化一个长度为 100 的桶(数组)
|
||
bucket := make([]*Entry, 100)
|
||
return &ArrayHashMap{bucket: bucket}
|
||
}
|
||
|
||
/* 哈希函数 */
|
||
func (a *ArrayHashMap) hashFunc(key int) int {
|
||
index := key % 100
|
||
return index
|
||
}
|
||
|
||
/* 查询操作 */
|
||
func (a *ArrayHashMap) get(key int) string {
|
||
index := a.hashFunc(key)
|
||
pair := a.bucket[index]
|
||
if pair == nil {
|
||
return "Not Found"
|
||
}
|
||
return pair.val
|
||
}
|
||
|
||
/* 添加操作 */
|
||
func (a *ArrayHashMap) put(key int, val string) {
|
||
pair := &Entry{key: key, val: val}
|
||
index := a.hashFunc(key)
|
||
a.bucket[index] = pair
|
||
}
|
||
|
||
/* 删除操作 */
|
||
func (a *ArrayHashMap) remove(key int) {
|
||
index := a.hashFunc(key)
|
||
// 置为空字符,代表删除
|
||
a.bucket[index] = nil
|
||
}
|
||
```
|
||
|
||
=== "JavaScript"
|
||
|
||
```js title="array_hash_map.js"
|
||
|
||
```
|
||
|
||
=== "TypeScript"
|
||
|
||
```typescript title="array_hash_map.ts"
|
||
|
||
```
|
||
|
||
=== "C"
|
||
|
||
```c title="array_hash_map.c"
|
||
|
||
```
|
||
|
||
=== "C#"
|
||
|
||
```csharp title="array_hash_map.cs"
|
||
|
||
```
|
||
|
||
## 哈希冲突
|
||
|
||
细心的同学可能会发现,**哈希函数 $f(x) = x \% 100$ 会在某些情况下失效**。具体地,当输入的 key 后两位相同时,哈希函数的计算结果也相同,指向同一个 value 。例如,分别查询两个学号 12836 和 20336 ,则有
|
||
|
||
$$
|
||
f(12836) = f(20336) = 36
|
||
$$
|
||
|
||
两个学号指向了同一个姓名,这明显是不对的,我们将这种现象称为「哈希冲突 Hash Collision」,其会严重影响查询的正确性。如何避免哈希冲突的问题将被留在下章讨论。
|
||
|
||
![hash_collision](hash_map.assets/hash_collision.png)
|
||
|
||
<p align="center"> Fig. 哈希冲突 </p>
|
||
|
||
综上所述,一个优秀的「哈希函数」应该具备以下特性:
|
||
|
||
- 尽量少地发生哈希冲突;
|
||
- 时间复杂度 $O(1)$ ,计算尽可能高效;
|
||
- 空间使用率高,即“键值对占用空间 / 哈希表总占用空间”尽可能大;
|