Given a sorted array of n integers that has been rotated an unknown number of times, write code to find an element in the array. You may assume that the array was originally sorted in increasing order. If there are more than one target elements in the array, return the smallest index.
Example1:
Input: arr = [15, 16, 19, 20, 25, 1, 3, 4, 5, 7, 10, 14], target = 5
Output: 8 (the index of 5 in the array)
We define the left boundary of the binary search as $l=0$ and the right boundary as $r=n-1$, where $n$ is the length of the array.
In each binary search process, we get the current midpoint $mid=(l+r)/2$.
If $nums[mid] > nums[r]$, it means that $[l,mid]$ is ordered. If $nums[l] \leq target \leq nums[mid]$, it means that $target$ is in $[l,mid]$, otherwise $target$ is in $[mid+1,r]$.
If $nums[mid] < nums[r]$, it means that $[mid+1,r]$ is ordered. If $nums[mid] < target \leq nums[r]$, it means that $target$ is in $[mid+1,r]$, otherwise $target$ is in $[l,mid]$.
If $nums[mid] = nums[r]$, it means that the elements $nums[mid]$ and $nums[r]$ are equal. At this time, we cannot determine which interval $target$ is in, we can only decrease $r$ by $1$.
After the binary search ends, if $nums[l] = target$, it means that the target value $target$ exists in the array, otherwise it does not exist.
Note that if initially $nums[l] = nums[r]$, we loop to decrease $r$ by $1$ until $nums[l] \neq nums[r]$.
The time complexity is approximately $O(\log n)$, and the space complexity is $O(1)$. Here, $n$ is the length of the array.