|
|
|
|
---
|
|
|
|
|
comments: true
|
|
|
|
|
---
|
|
|
|
|
|
|
|
|
|
# 10.2. 二分查找边界
|
|
|
|
|
|
|
|
|
|
上一节规定目标元素在数组中是唯一的。如果目标元素在数组中多次出现,上节介绍的方法只能保证返回其中一个目标元素的索引,**而无法确定该索引的左边和右边还有多少目标元素**。
|
|
|
|
|
|
|
|
|
|
为了查找最左一个 `target` ,我们可以先进行二分查找,找到任意一个目标元素,**再加一个向左遍历的线性查找**,找到最左的 `target` 返回即可。然而,由于加入了线性查找,这个方法的时间复杂度可能会劣化至 $O(n)$ 。
|
|
|
|
|
|
|
|
|
|
![线性查找最左元素](binary_search_edge.assets/binary_search_left_edge_naive.png)
|
|
|
|
|
|
|
|
|
|
<p align="center"> Fig. 线性查找最左元素 </p>
|
|
|
|
|
|
|
|
|
|
## 10.2.1. 查找最左一个元素
|
|
|
|
|
|
|
|
|
|
!!! question
|
|
|
|
|
|
|
|
|
|
给定一个长度为 $n$ 的有序数组 `nums` 。请查找并返回元素 `target` 在该数组中首次出现的索引。若数组中不包含该元素,则返回 $-1$ 。数组可能包含重复元素。
|
|
|
|
|
|
|
|
|
|
实际上,我们可以仅通过二分查找解决以上问题。方法的整体框架不变,先计算中点索引 `m` ,再判断 `target` 和 `nums[m]` 大小关系:
|
|
|
|
|
|
|
|
|
|
- 当 `nums[m] < target` 或 `nums[m] > target` 时,说明还没有找到 `target` ,因此采取与上节代码相同的缩小区间操作。
|
|
|
|
|
- 当 `nums[m] == target` 时,说明找到了一个目标元素,此时应该如何缩小区间?
|
|
|
|
|
|
|
|
|
|
对于该情况,**我们可以将查找目标想象为 `leftarget`**,其中 `leftarget` 表示从右到左首个小于 `target` 的元素。具体来说:
|
|
|
|
|
|
|
|
|
|
- 当 `nums[m] == target` 时,说明 `leftarget` 在区间 `[i, m - 1]` 中,因此采用 `j = m - 1` 来缩小区间,**从而使指针 `j` 向 `leftarget` 收缩靠近**。
|
|
|
|
|
- 二分查找完成后,`i` 指向最左一个 `target` ,`j` 指向 `leftarget` ,因此最终返回索引 `i` 即可。
|
|
|
|
|
|
|
|
|
|
=== "<1>"
|
|
|
|
|
![二分查找最左元素的步骤](binary_search_edge.assets/binary_search_left_edge_step1.png)
|
|
|
|
|
|
|
|
|
|
=== "<2>"
|
|
|
|
|
![binary_search_left_edge_step2](binary_search_edge.assets/binary_search_left_edge_step2.png)
|
|
|
|
|
|
|
|
|
|
=== "<3>"
|
|
|
|
|
![binary_search_left_edge_step3](binary_search_edge.assets/binary_search_left_edge_step3.png)
|
|
|
|
|
|
|
|
|
|
=== "<4>"
|
|
|
|
|
![binary_search_left_edge_step4](binary_search_edge.assets/binary_search_left_edge_step4.png)
|
|
|
|
|
|
|
|
|
|
=== "<5>"
|
|
|
|
|
![binary_search_left_edge_step5](binary_search_edge.assets/binary_search_left_edge_step5.png)
|
|
|
|
|
|
|
|
|
|
=== "<6>"
|
|
|
|
|
![binary_search_left_edge_step6](binary_search_edge.assets/binary_search_left_edge_step6.png)
|
|
|
|
|
|
|
|
|
|
=== "<7>"
|
|
|
|
|
![binary_search_left_edge_step7](binary_search_edge.assets/binary_search_left_edge_step7.png)
|
|
|
|
|
|
|
|
|
|
=== "<8>"
|
|
|
|
|
![binary_search_left_edge_step8](binary_search_edge.assets/binary_search_left_edge_step8.png)
|
|
|
|
|
|
|
|
|
|
注意,数组可能不包含目标元素 `target` 。因此在函数返回前,我们需要先判断 `nums[i]` 与 `target` 是否相等。另外,当 `target` 大于数组中的所有元素时,索引 `i` 会越界,因此也需要额外判断。
|
|
|
|
|
|
|
|
|
|
=== "Java"
|
|
|
|
|
|
|
|
|
|
```java title="binary_search_edge.java"
|
|
|
|
|
/* 二分查找最左一个元素 */
|
|
|
|
|
int binarySearchLeftEdge(int[] nums, int target) {
|
|
|
|
|
int i = 0, j = nums.length - 1; // 初始化双闭区间 [0, n-1]
|
|
|
|
|
while (i <= j) {
|
|
|
|
|
int m = i + (j - i) / 2; // 计算中点索引 m
|
|
|
|
|
if (nums[m] < target)
|
|
|
|
|
i = m + 1; // target 在区间 [m+1, j] 中
|
|
|
|
|
else if (nums[m] > target)
|
|
|
|
|
j = m - 1; // target 在区间 [i, m-1] 中
|
|
|
|
|
else
|
|
|
|
|
j = m - 1; // 首个小于 target 的元素在区间 [i, m-1] 中
|
|
|
|
|
}
|
|
|
|
|
if (i == nums.length || nums[i] != target)
|
|
|
|
|
return -1; // 未找到目标元素,返回 -1
|
|
|
|
|
return i;
|
|
|
|
|
}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C++"
|
|
|
|
|
|
|
|
|
|
```cpp title="binary_search_edge.cpp"
|
|
|
|
|
/* 二分查找最左一个元素 */
|
|
|
|
|
int binarySearchLeftEdge(vector<int> &nums, int target) {
|
|
|
|
|
int i = 0, j = nums.size() - 1; // 初始化双闭区间 [0, n-1]
|
|
|
|
|
while (i <= j) {
|
|
|
|
|
int m = i + (j - i) / 2; // 计算中点索引 m
|
|
|
|
|
if (nums[m] < target)
|
|
|
|
|
i = m + 1; // target 在区间 [m+1, j] 中
|
|
|
|
|
else if (nums[m] > target)
|
|
|
|
|
j = m - 1; // target 在区间 [i, m-1] 中
|
|
|
|
|
else
|
|
|
|
|
j = m - 1; // 首个小于 target 的元素在区间 [i, m-1] 中
|
|
|
|
|
}
|
|
|
|
|
if (i == nums.size() || nums[i] != target)
|
|
|
|
|
return -1; // 未找到目标元素,返回 -1
|
|
|
|
|
return i;
|
|
|
|
|
}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Python"
|
|
|
|
|
|
|
|
|
|
```python title="binary_search_edge.py"
|
|
|
|
|
def binary_search_left_edge(nums: list[int], target: int) -> int:
|
|
|
|
|
"""二分查找最左一个元素"""
|
|
|
|
|
# 初始化双闭区间 [0, n-1] ,即 i, j 分别指向数组首元素、尾元素
|
|
|
|
|
i, j = 0, len(nums) - 1
|
|
|
|
|
while i <= j:
|
|
|
|
|
m = (i + j) // 2 # 计算中点索引 m
|
|
|
|
|
if nums[m] < target:
|
|
|
|
|
i = m + 1 # target 在区间 [m+1, j] 中
|
|
|
|
|
elif nums[m] > target:
|
|
|
|
|
j = m - 1 # target 在区间 [i, m-1] 中
|
|
|
|
|
else:
|
|
|
|
|
j = m - 1 # 首个小于 target 的元素在区间 [i, m-1] 中
|
|
|
|
|
if i == len(nums) or nums[i] != target:
|
|
|
|
|
return -1 # 未找到目标元素,返回 -1
|
|
|
|
|
return i
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Go"
|
|
|
|
|
|
|
|
|
|
```go title="binary_search_edge.go"
|
|
|
|
|
[class]{}-[func]{binarySearchLeftEdge}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "JavaScript"
|
|
|
|
|
|
|
|
|
|
```javascript title="binary_search_edge.js"
|
|
|
|
|
[class]{}-[func]{binarySearchLeftEdge}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "TypeScript"
|
|
|
|
|
|
|
|
|
|
```typescript title="binary_search_edge.ts"
|
|
|
|
|
[class]{}-[func]{binarySearchLeftEdge}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C"
|
|
|
|
|
|
|
|
|
|
```c title="binary_search_edge.c"
|
|
|
|
|
[class]{}-[func]{binarySearchLeftEdge}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C#"
|
|
|
|
|
|
|
|
|
|
```csharp title="binary_search_edge.cs"
|
|
|
|
|
[class]{binary_search_edge}-[func]{binarySearchLeftEdge}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Swift"
|
|
|
|
|
|
|
|
|
|
```swift title="binary_search_edge.swift"
|
|
|
|
|
[class]{}-[func]{binarySearchLeftEdge}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Zig"
|
|
|
|
|
|
|
|
|
|
```zig title="binary_search_edge.zig"
|
|
|
|
|
[class]{}-[func]{binarySearchLeftEdge}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
## 10.2.2. 查找最右一个元素
|
|
|
|
|
|
|
|
|
|
类似地,我们也可以二分查找最右一个元素。设首个大于 `target` 的元素为 `rightarget` 。
|
|
|
|
|
|
|
|
|
|
- 当 `nums[m] == target` 时,说明 `rightarget` 在区间 `[m + 1, j]` 中,因此执行 `i = m + 1` 将搜索区间向右收缩。
|
|
|
|
|
- 完成二分后,`i` 指向 `rightarget` ,`j` 指向最右一个 `target` ,因此最终返回索引 `j` 即可。
|
|
|
|
|
|
|
|
|
|
=== "Java"
|
|
|
|
|
|
|
|
|
|
```java title="binary_search_edge.java"
|
|
|
|
|
/* 二分查找最右一个元素 */
|
|
|
|
|
int binarySearchRightEdge(int[] nums, int target) {
|
|
|
|
|
int i = 0, j = nums.length - 1; // 初始化双闭区间 [0, n-1]
|
|
|
|
|
while (i <= j) {
|
|
|
|
|
int m = i + (j - i) / 2; // 计算中点索引 m
|
|
|
|
|
if (nums[m] < target)
|
|
|
|
|
i = m + 1; // target 在区间 [m+1, j] 中
|
|
|
|
|
else if (nums[m] > target)
|
|
|
|
|
j = m - 1; // target 在区间 [i, m-1] 中
|
|
|
|
|
else
|
|
|
|
|
i = m + 1; // 首个大于 target 的元素在区间 [m+1, j] 中
|
|
|
|
|
}
|
|
|
|
|
if (j < 0 || nums[j] != target)
|
|
|
|
|
return -1; // 未找到目标元素,返回 -1
|
|
|
|
|
return j;
|
|
|
|
|
}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C++"
|
|
|
|
|
|
|
|
|
|
```cpp title="binary_search_edge.cpp"
|
|
|
|
|
/* 二分查找最右一个元素 */
|
|
|
|
|
int binarySearchRightEdge(vector<int> &nums, int target) {
|
|
|
|
|
int i = 0, j = nums.size() - 1; // 初始化双闭区间 [0, n-1]
|
|
|
|
|
while (i <= j) {
|
|
|
|
|
int m = i + (j - i) / 2; // 计算中点索引 m
|
|
|
|
|
if (nums[m] < target)
|
|
|
|
|
i = m + 1; // target 在区间 [m+1, j] 中
|
|
|
|
|
else if (nums[m] > target)
|
|
|
|
|
j = m - 1; // target 在区间 [i, m-1] 中
|
|
|
|
|
else
|
|
|
|
|
i = m + 1; // 首个大于 target 的元素在区间 [m+1, j] 中
|
|
|
|
|
}
|
|
|
|
|
if (j < 0 || nums[j] != target)
|
|
|
|
|
return -1; // 未找到目标元素,返回 -1
|
|
|
|
|
return j;
|
|
|
|
|
}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Python"
|
|
|
|
|
|
|
|
|
|
```python title="binary_search_edge.py"
|
|
|
|
|
def binary_search_right_edge(nums: list[int], target: int) -> int:
|
|
|
|
|
"""二分查找最右一个元素"""
|
|
|
|
|
# 初始化双闭区间 [0, n-1] ,即 i, j 分别指向数组首元素、尾元素
|
|
|
|
|
i, j = 0, len(nums) - 1
|
|
|
|
|
while i <= j:
|
|
|
|
|
m = (i + j) // 2 # 计算中点索引 m
|
|
|
|
|
if nums[m] < target:
|
|
|
|
|
i = m + 1 # target 在区间 [m+1, j] 中
|
|
|
|
|
elif nums[m] > target:
|
|
|
|
|
j = m - 1 # target 在区间 [i, m-1] 中
|
|
|
|
|
else:
|
|
|
|
|
i = m + 1 # 首个大于 target 的元素在区间 [m+1, j] 中
|
|
|
|
|
if j == len(nums) or nums[j] != target:
|
|
|
|
|
return -1 # 未找到目标元素,返回 -1
|
|
|
|
|
return j
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Go"
|
|
|
|
|
|
|
|
|
|
```go title="binary_search_edge.go"
|
|
|
|
|
[class]{}-[func]{binarySearchRightEdge}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "JavaScript"
|
|
|
|
|
|
|
|
|
|
```javascript title="binary_search_edge.js"
|
|
|
|
|
[class]{}-[func]{binarySearchRightEdge}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "TypeScript"
|
|
|
|
|
|
|
|
|
|
```typescript title="binary_search_edge.ts"
|
|
|
|
|
[class]{}-[func]{binarySearchRightEdge}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C"
|
|
|
|
|
|
|
|
|
|
```c title="binary_search_edge.c"
|
|
|
|
|
[class]{}-[func]{binarySearchRightEdge}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "C#"
|
|
|
|
|
|
|
|
|
|
```csharp title="binary_search_edge.cs"
|
|
|
|
|
[class]{binary_search_edge}-[func]{binarySearchRightEdge}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Swift"
|
|
|
|
|
|
|
|
|
|
```swift title="binary_search_edge.swift"
|
|
|
|
|
[class]{}-[func]{binarySearchRightEdge}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
=== "Zig"
|
|
|
|
|
|
|
|
|
|
```zig title="binary_search_edge.zig"
|
|
|
|
|
[class]{}-[func]{binarySearchRightEdge}
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
观察下图,搜索最右元素时指针 `j` 起到了搜索最左元素时指针 `i` 的作用,反之亦然。本质上看,**搜索最左元素和最右元素的实现是镜像对称的**。
|
|
|
|
|
|
|
|
|
|
![二分查找最左元素和最右元素](binary_search_edge.assets/binary_search_left_right_edge.png)
|
|
|
|
|
|
|
|
|
|
<p align="center"> Fig. 二分查找最左元素和最右元素 </p>
|
|
|
|
|
|
|
|
|
|
!!! tip
|
|
|
|
|
|
|
|
|
|
以上代码采取的都是“双闭区间”写法。有兴趣的读者可以自行实现“左闭右开”写法。
|