通过索引过滤Numpy值 修复数组比较报错获取预期输出

错误原因 #

1. 语法错误：相等判断误用了赋值运算符=，应该用比较运算符==，同时np.where的参数传递不符合语法要求，触发positional argument follows keyword argument报错
2. 逻辑错误：两个unix时间数组长度不一致，直接逐位比较会触发numpy广播机制，无法正确匹配到两个数组共有的时间点对应的索引

修复后代码 #

import numpy as np 

a=np.array(["Short","Long","Short","Long","Short","Long"])
b=np.array(["Short","Long","Short","Long"])
c=np.array(["Short","Long","Short","Long","Long","Long"])

unix_a=np.array([1624580882,1624584458,1624589467,1624592213,1624595336,1624596349])
unix_b=np.array([1624580882,1624584458,1624595336,1624596349])
unix_c=np.array([1624580882,1624584464,1624589495,1624592238,1624595350,1624596380])

def Compare(val,val2, unix_val, unix_val2):
    # 取两个unix数组的交集，同时返回对应两个数组的索引
    common_unix, idx_unix_val, idx_unix_val2 = np.intersect1d(unix_val, unix_val2, return_indices=True)
    # 按共同时间点过滤两个val数组
    filtered_val = val[idx_unix_val]
    filtered_val2 = val2[idx_unix_val2]
    # 匹配val相等的位置，返回原始unix_val对应的索引
    equal_mask = filtered_val == filtered_val2
    return idx_unix_val[equal_mask].tolist()

print(Compare(a, b, unix_a, unix_b))
print(Compare(a, c, unix_a, unix_c))

运行输出 #

[0, 1, 4, 5]
[0]

完全符合预期要求。

内容的提问来源于stack exchange，提问作者georgehere