Python 循环内导入比文件级导入慢约 3 倍

import time
import sys

def benchmark_import_outside():
    """对在循环外导入 math 模块进行基准测试。"""
    import math  # 在循环外导入一次

    start = time.time()
    for _ in range(1_000_000):
        x = math.sqrt(100)  # 使用已导入的模块
    end = time.time()

    return end - start

def benchmark_import_inside():
    """对在循环内导入 math 模块进行基准测试。"""
    start = time.time()
    for _ in range(1_000_000):
        import math  # 在循环内导入（不良实践）
        x = math.sqrt(100)
    end = time.time()

    return end - start

def clear_math_from_cache():
    """从 sys.modules 中清除 math 模块以强制重新加载。"""
    if 'math' in sys.modules:
        del sys.modules['math']

def main():
    # 基准测试前确保 math 未被缓存
    clear_math_from_cache()

    # 对外部导入进行基准测试
    time_outside = benchmark_import_outside()
    print(f"Time (import outside loop): {time_outside:.4f} sec")

    # 从缓存中清除 math 以确保公平比较
    clear_math_from_cache()

    # 对内部导入进行基准测试
    time_inside = benchmark_import_inside()
    print(f"Time (import inside loop): {time_inside:.4f} sec")

    # 计算并打印性能差异
    ratio = time_inside / time_outside
    print(f"Import inside loop is {ratio:.0f}x slower than importing outside.")

if __name__ == "__main__":
    main()

Time (import outside loop): 0.0786 sec
Time (import inside loop): 0.2162 sec
Import inside loop is 3x slower than importing outside.