How to implement Python search algorithm

The search algorithm is used to retrieve whether there is a given data (keyword) in the sequence data (population). Commonly used search algorithms are:

• Linear search: Linear search is also called It is a sequential search, used to search in unordered numerical columns.

• Binary search: Binary search is also called half search, and its algorithm is used for ordered sequences.

• Interpolation search: Interpolation search is an improvement on the binary search algorithm.

• Blocked search: Also known as index sequential search, it is an improved version of linear search.

• Tree table search: Tree table search can be divided into binary search tree and balanced binary tree search.

• Hash search: Hash search can directly find the required data through keywords.

Because tree table search and hash search require a lot of space, they will not be explained in this article. This article provides a comprehensive overview of search algorithms beyond tree-based and hash-based approaches. It analyzes the strengths and weaknesses of each algorithm and proposes corresponding optimization strategies.

1. Linear search

Sequential search, also known as linear search, is an algorithm based on primitive, exhaustive, and brute force search. It is easy to understand and the coding implementation is simple. If the amount of data processed is large, its performance may be lower due to the relatively simple idea of ​​the algorithm and the lack of optimal design of the algorithm.

Linear search idea:

• Scan each data in the original list one by one from beginning to end, and compare it with the given keyword.

• If the comparison is equal, the search is successful.

• When the scan is completed and data equal to the given keyword is still not found, the search failure is declared.

According to the description of the linear search algorithm, it is easy to code and implement:

'''
线性查找算法
参数：
    nums: 序列
    key:关键字
返回值：
    关键字在序列中的位置
    如果没有，则返回 -1
'''
def line_find(nums, key):
    for i in range(len(nums)):
        if nums[i] == key:
            return i
    return -1
'''
测试线性算法
'''
if __name__ == "__main__":
    nums = [4, 1, 8, 10, 3, 5]
    key = int(input("请输入要查找的关键字："))
    pos = line_find(nums, key)
    print("关键字 {0} 在数列的第 {1} 位置".format(key, pos))
'''
输出结果：
请输入要查找的关键字：3
关键字 3 在数列的 4 位置
'''

Average time complexity analysis of the linear search algorithm.

1. Best luck situation: If the keyword you want to find happens to be at the 1st position of the sequence, you only need to search it once.

For example, search for the keyword 4 in the sequence =[4,1,8,10,3,5].

Only needs to be searched once.

2. The worst-luck situation: the keyword is not found until the end of the sequence is scanned.

Such as searching for the keyword 5 in the sequence =[4,1,8,10,3,5].

The number of searches required is equal to the length of the sequence, which is 6 times here.

3. Good or bad luck: If the keyword to be found is somewhere in the middle of the sequence, the probability of finding it is 1/n.

n is the length of the sequence.

The average number of searches for linear search should = (1 n)/2. This sentence is rewritten as: Its time complexity is O(n).

Constants are ignored in big O notation.

The worst case scenario for linear search is: after scanning the entire array, there is no keyword to be found.

For example, find whether the keyword 12 exists in the sequence =[4,1,8,10,3,5].

After scanning 6 times, I failed! !

Improved Linear Search Algorithm

The Linear Search algorithm can be optimized accordingly. Such as setting up an "outpost". The so-called "outpost" is to insert the keyword to be searched for at the end of the sequence before searching.

def line_find_(nums, key):
    i = 0
    while nums[i] != key:
        i += 1
    return -1 if i == len(nums)-1 else i

'''
测试线性算法
'''
if __name__ == "__main__":
    nums = [4, 1, 8, 10, 3, 5]
    key = int(input("请输入要查找的关键字："))
    # 查找之前，先把关键字存储到列到的尾部
    nums.append(key)
    pos = line_find_(nums, key)
    print("关键字 {0} 在数列的第 {1} 位置".format(key, pos))

The time complexity of the linear search algorithm optimized with "Outpost" has not changed, O(n). In other words, judging from the 2 code, there are not many changes.

But from the perspective of the actual running of the code, the 2 option reduces the number of if instructions and also reduces the number of compiled instructions, thus reducing CPUThe number of times the instruction is executed, this kind of optimization is a micro-optimization, not an optimization of the essence of the algorithm.

Code written using computer programming language is pseudo-instruction code.

The compiled instruction code is called CPU instruction set.

One optimization solution is to reduce the compiled instruction set.

2. Binary search

Ordered search means that the data being searched must be arranged in a certain order, and binary search is an ordered search. For example, if you search for the keyword 4 in the sequence = [4,1,8,10,3,5,12], the factor sequence is not ordered, so binary search cannot be used. If you want to use the binary search algorithm, you need to first Sort a sequence of numbers.

Binary search uses the idea of ​​bisection (halving) algorithm. There are 2 key information in the binary search algorithm that need to be obtained at any time:

• One is the middle position of the sequence mid_pos.

• One is the middle value mid_val of the sequence.

Now learn about the algorithm process of binary search by searching for the keyword 8 in the sequence nums=[1,3,4,5,8,10,12].

Before performing a binary search, first define 2 position (pointer) variables:

• The left pointer l_idx initially points to the leftmost number of the sequence.

• The right pointer r_idx initially points to the rightmost number of the array.

第 1 步：通过左、右指针的当前位置计算出数列的中间位置 mid_pos=3，并根据 mid_pos 的值找出数列中间位置所对应的值 mid_val=nums[mid_pos] 是 5。

二分查找算法的核心就是要找出数列中间位置的值。

第 2 步：把数列中间位置的值和给定的关键字相比较。这里关键字是 8，中间位置的值是 5，显然 8 是大于 5，因为数列是有序的，自然会想到没有必要再与数列中 5 之前的数字比较，而是专心和 5 之后的数字比较。

一次比较后再次查找的数列范围缩小了一半。这也是二分算法的由来。

第 3 步：根据比较结果，调整数列的大小，这里的大小调整不是物理结构上调整，而是逻辑上调整，调整后原数列没有变化。也就是通过修改左指针或右指针的位置，从逻辑上改变数列大小。调整后的数列如下图。

二分查找算法中数列的范围由左指针到右指针的长度决定。

第 4 步：重复上述步骤，至到找到或找不到为止。

编码实现二分查找算法

'''
二分查找算法
'''
def binary_find(nums, key):
    # 初始左指针
    l_idx = 0
    # 初始在指针
    r_ldx = len(nums) - 1
    while l_idx <= r_ldx:
        # 计算出中间位置
        mid_pos = (r_ldx + l_idx) // 2
        # 计算中间位置的值
        mid_val = nums[mid_pos]
        # 与关键字比较
        if mid_val == key:
            # 出口一：比较相等，有此关键字，返回关键字所在位置
            return mid_pos
        elif mid_val > key:
            # 说明查找范围应该缩少在原数的左边
            r_ldx = mid_pos - 1
        else:
            l_idx = mid_pos + 1
    # 出口二：没有查找到给定关键字
    return -1

&#39;&#39;&#39;
测试二分查找
&#39;&#39;&#39;
if __name__ == "__main__":
    nums = [1, 3, 4, 5, 8, 10, 12]
    key = 3
    pos = binary_find(nums, key)
    print(pos)

通过前面对二分算法流程的分析，可知二分查找的子问题和原始问题是同一个逻辑，所以可以使用递归实现：

'''
递归实现二分查找
'''
def binary_find_dg(nums, key, l_idx, r_ldx):
    if l_idx > r_ldx:
        # 出口一：没有查找到给定关键字
        return -1
    # 计算出中间位置
    mid_pos = (r_ldx + l_idx) // 2
    # 计算中间位置的值
    mid_val = nums[mid_pos]
    # 与关键字比较
    if mid_val == key:
        # 出口二：比较相等，有此关键字，返回关键字所在位置
        return mid_pos
    elif mid_val > key:
        # 说明查找范围应该缩少在原数的左边
        r_ldx = mid_pos - 1
    else:
        l_idx = mid_pos + 1
    return binary_find_dg(nums, key, l_idx, r_ldx)
'''
测试二分查找
'''
if __name__ == "__main__":
    nums = [1, 3, 4, 5, 8, 10, 12]
    key = 8
    pos = binary_find_dg(nums, key,0,len(nums)-1)
    print(pos)

二分查找性能分析：

二分查找的过程用树形结构描述会更直观，当搜索完毕后，绘制出来树结构是一棵二叉树。

1.如上述代码执行过程中，先找到数列中的中间数字 5，然后以 5 为根节点构建唯一结点树。

2.5 和关键字 8 比较后，再在以数字 5 为分界线的右边数列中找到中间数字10，树形结构会变成下图所示。

3.10 和关键字 8比较后，再在10 的左边查找。

查找到8 后，意味着二分查找已经找到结果，只需要 3 次就能查找到最终结果。

从二叉树的结构上可以直观得到结论：二分查找关键字的次数由关键字在二叉树结构中的深度决定。

4.上述是查找给定的数字8，为了能查找到数列中的任意一个数字，最终完整的树结构应该如下图所示。

很明显，树结构是标准的二叉树。从树结构上可以看出，无论查找任何数字，最小是 1 次，如查找数字 5，最多也只需要 3 次，比线性查找要快很多。

根据二叉树的特性，结点个数为 n 的树的深度为 h=log2(n+1)，所以二分查找算法的大 O 表示的时间复杂度为 O(logn)，是对数级别的时间度。

当对长度为1000的数列进行二分查找时，所需次数最多只要 10 次，二分查找算法的效率显然是高效的。

然而，二分查找算法在实行之前需要对数列进行排序，因此前面所述的时间复杂度并未包含排序所需的时间。所以，二分查找一般适合数字变化稳定的有序数列。

3. 插值查找

插值查找本质是二分查找，插值查找对二分查找算法中查找中间位置的计算逻辑进行了改进。

原生二分查找算法中计算中间位置的逻辑：中间位置等于左指针位置加上右指针位置然后除以 2。

    # 计算中间位置
    mid_pos = (r_ldx + l_idx) // 2

插值算法计算中间位置逻辑如下所示：

key 为要查找的关键字！！

# 插值算法中计算中间位置
mid_pos = l_idx + (key - nums[l_idx]) // (nums[r_idx] - nums[l_idx]) * (r_idx - l_idx)

编码实现插值查找：

# 插值查找基于二分法，只是mid计算方法不同
def binary_search(nums, key):
    l_idx = 0
    r_idx = len(nums) - 1
    old_mid = -1
    mid_pos = None
    while l_idx < r_idx and nums[0] <= key and nums[r_idx] >= key and old_mid != mid_pos:
        # 中间位置计算
        mid_pos = l_idx + (key - nums[l_idx]) // (nums[r_idx] - nums[l_idx]) * (r_idx - l_idx)
        old_mid = mid_pos
        if nums[mid_pos] == key:
            return "index is {}, target value is {}".format(mid_pos, nums[mid_pos])
            # 此时目标值在中间值右边，更新左边界位置
        elif nums[mid_pos] < key:
            l_idx = mid_pos + 1
        # 此时目标值在中间值左边，更新右边界位置
        elif nums[mid_pos] > key:
            r_idx = mid_pos - 1
    return "Not find"

li =[1, 3, 4, 5, 8, 10, 12]
print(binary_search(li, 6))

插值算法的中间位置计算时，对中间位置的计算有可能多次计算的结果是一样的，此时可以认为查找失败。

插值算法的性能介于线性查找和二分查找之间。

如果序列具有较大数量的均匀分布的数字，插值查找算法的平均执行效率要比二分查找好得多。如果数据在数列中分布不均匀，插值算法并不是最优选择。

4. 分块查找

分块查找类似于数据库中的索引查询，所以分块查找也称为索引查找。其算法的核心还是线性查找。

现有原始数列 nums=[5,1,9,11,23,16,12,18,24,32,29,25]，需要查找关键字11 是否存在。

第 1 步：使用分块查找之前，先要对原始数列按区域分成多个块。至于分成多少块，可根据实际情况自行定义。分块时有一个要求，前一个块中的最大值必须小于后一个块的最小值。

块内部无序，但要保持整个数列按块有序。

分块查找要求原始数列从整体上具有升序或降序趋势，如果数列的分布不具有趋向性，如果仍然想使用分块查找，则需要进行分块有序调整。

第 2 步：根据分块信息，建立索引表。索引表至少应该有 2 个字段，每一块中的最大值数字以及每一块的起始地址。显然索引表中的数字是有序的。

第 3 步：查找给定关键字时，先查找索引表，查询关键字应该在那个块中。如查询关键字 29，可知应该在第三块中，然后根据索引表中所提供的第三块的地址信息，再进入第三块数列，按线性匹配算法查找29 具体位置。

编码实现分块查找：

先编码实现根据分块数量、创建索引表，这里使用二维列表保存储索引表中的信息。

'''
分块：建立索引表
参数：
    nums 原始数列
    blocks 块大小
'''
def create_index_table(nums, blocks):
    # 索引表使用列表保存
    index_table = []
    # 每一块的数量
    n = len(nums) // blocks
    for i in range(0, len(nums), n):
        # 索引表中的每一行记录
        tmp_lst = []
        # 最大值
        tmp_lst.append(max(nums[i:i + n-1]))
        # 起始地址
        tmp_lst.append(i)
        # 终止地址
        tmp_lst.append(i + n - 1)
        # 添加到索引表中
        index_table.append(tmp_lst)
    return index_table
'''
测试分块
'''
nums = [5, 1, 9, 11, 23, 16, 12, 18, 24, 32, 29, 25]
it = create_index_table(nums, 3)
print(it)
'''
输出结果：
[[11, 0, 3], [23, 4, 7], [32, 8, 11]]
'''

代码执行后，输出结果和分析的结果一样。

以上代码仅对整体趋势有序的数列进行分块。如果整体没有向有序趋势发展，则需要提供适当的块排序计划，有兴趣的人可以自行完成。

如上代码仅为说明分块查找算法。

分块查找的完整代码：

'''
分块：建立索引表
参数：
    nums 原始数列
    blocks 块大小
'''
def create_index_table(nums, blocks):
    # 索引表使用列表保存
    index_table = []
    # 每一块的数量
    n = len(nums) // blocks
    for i in range(0, len(nums), n):
        tmp_lst = []
        tmp_lst.append(max(nums[i:i + n - 1]))
        tmp_lst.append(i)
        tmp_lst.append(i + n - 1)
        index_table.append(tmp_lst)
    return index_table

'''
使用线性查找算法在对应的块中查找
'''
def lind_find(nums, start, end):
    for i in range(start, end):
        if key == nums[i]:
            return i
            break
    return -1

'''
测试分块
'''
nums = [5, 1, 9, 11, 23, 16, 12, 18, 24, 32, 29, 25]
key = 16
# 索引表
it = create_index_table(nums, 3)
# 索引表的记录编号
pos = -1
# 在索引表中查询
for n in range(len(it) - 1):
    # 是不是在第一块中
    if key <= it[0][0]:
        pos = 0
    # 其它块中
    if it[n][0] < key <= it[n + 1][0]:
        pos = n + 1
        break
if pos == -1:
    print("{0} 在 {1} 数列中不存在".format(key, nums))
else:
    idx = lind_find(nums, it[pos][1], it[pos][2] + 1)
    if idx != -1:
        print("{0} 在 {1} 数列的 {2} 位置".format(key, nums, idx))
    else:
        print("{0} 在 {1} 数列中不存在".format(key, nums))
&#39;&#39;&#39;
输出结果
16 在 [5, 1, 9, 11, 23, 16, 12, 18, 24, 32, 29, 25] 数列的第 5 位置
&#39;&#39;&#39;

分块查找对于整体趋向有序的数列，其查找性能较好。如果原始数列没有整体有序性，就需要使用块排序算法，其时间复杂度没有二分查找算法好。

建立索引表是分块查找所必需的，但这会增加额外的存储空间，因此其空间复杂度较高。其优于二分的地方在于只需要对原始数列进行部分排序。本质还是以线性查找为主。

The above is the detailed content of How to implement Python search algorithm. For more information, please follow other related articles on the PHP Chinese website!