今天分享 18 个 Python 坏习惯,这些坏习惯会暴露开发者在 Python 方面经验不足。通过摒弃这些习惯并以 Pythonic 的方式编写代码,可以提高你的代码质量,给看代码的人留下好印象。 喜欢本文记得收藏、关注、点赞。
【注】文末加入技术交流群,学习不再孤单
李宏毅《机器学习》国语课程(2022)来了
有人把吴恩达老师的机器学习和深度学习做成了中文版
上瘾了,最近又给公司撸了一个可视化大屏(附源码)
如此优雅,4款 Python 自动数据分析神器真香啊
梳理半月有余,精心准备了17张知识思维导图,这次要讲清统计学
年终汇总:20份可视化大屏模板,直接套用真香(文末附源码)
坏的做法:
def manual_str_formatting(name, subscribers):
if subscribers > 100000:
print("Wow " + name + "! you have " + str(subscribers) + " subscribers!")
else:
print("Lol " + name + " that's not many subs")
好的做法是使用 f-string,而且效率会更高:
def manual_str_formatting(name, subscribers):
# better
if subscribers > 100000:
print(f"Wow {name}! you have {subscribers} subscribers!")
else:
print(f"Lol {name} that's not many subs")
坏的做法:
def finally_instead_of_context_manager(host, port):
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
try:
s.connect((host, port))
s.sendall(b'Hello, world')
finally:
s.close()
好的做法是使用上下文管理器,即使发生异常,也会关闭 socket::
def finally_instead_of_context_manager(host, port):
# close even if exception
with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
s.connect((host, port))
s.sendall(b'Hello, world')
坏的做法:
def manually_calling_close_on_a_file(filename):
f = open(filename, "w")
f.write("hello!\n")
f.close()
好的做法是使用上下文管理器,即使发生异常,也会自动关闭文件,凡是有上下文管理器的,都应该首先采用:
def manually_calling_close_on_a_file(filename):
with open(filename) as f:
f.write("hello!\n")
# close automatic, even if exception
坏的做法:
def bare_except():
while True:
try:
s = input("Input a number: ")
x = int(s)
break
except: # oops! can't CTRL-C to exit
print("Not a number, try again")
这样会捕捉所有异常,导致按下 CTRL-C 程序都不会终止,好的做法是
def bare_except():
while True:
try:
s = input("Input a number: ")
x = int(s)
break
except Exception: # 比这更好的是用 ValueError
print("Not a number, try again")
如果函数参数使用可变对象,那么下次调用时可能会产生非预期结果,坏的做法
def mutable_default_arguments():
def append(n, l=[]):
l.append(n)
return l
l1 = append(0) # [0]
l2 = append(1) # [0, 1]
好的做法:
def mutable_default_arguments():
def append(n, l=None):
if l is None:
l = []
l.append(n)
return l
l1 = append(0) # [0]
l2 = append(1) # [1]
坏的做法
squares = {}
for i in range(10):
squares[i] = i * i
好的做法
odd_squares = {i: i * i for i in range(10)}
推导式虽然好用,但是不可以牺牲可读性,坏的做法
c = [
sum(a[n * i + k] * b[n * k + j] for k in range(n))
for i in range(n)
for j in range(n)
]
好的做法:
c = []
for i in range(n):
for j in range(n):
ij_entry = sum(a[n * i + k] * b[n * k + j] for k in range(n))
c.append(ij_entry)
坏的做法
def checking_type_equality():
Point = namedtuple('Point', ['x', 'y'])
p = Point(1, 2)
if type(p) == tuple:
print("it's a tuple")
else:
print("it's not a tuple")
好的做法
def checking_type_equality():
Point = namedtuple('Point', ['x', 'y'])
p = Point(1, 2)
# probably meant to check if is instance of tuple
if isinstance(p, tuple):
print("it's a tuple")
else:
print("it's not a tuple")
坏的做法
def equality_for_singletons(x):
if x == None:
pass
if x == True:
pass
if x == False:
pass
好的做法
def equality_for_singletons(x):
# better
if x is None:
pass
if x is True:
pass
if x is False:
pass
坏的做法
def checking_bool_or_len(x):
if bool(x):
pass
if len(x) != 0:
pass
好的做法
def checking_bool_or_len(x):
# usually equivalent to
if x:
pass
坏的做法
def range_len_pattern():
a = [1, 2, 3]
for i in range(len(a)):
v = a[i]
...
b = [4, 5, 6]
for i in range(len(b)):
av = a[i]
bv = b[i]
...
好的做法
def range_len_pattern():
a = [1, 2, 3]
# instead
for v in a:
...
# or if you wanted the index
for i, v in enumerate(a):
...
# instead use zip
for av, bv in zip(a, b):
...
坏的做法
def not_using_dict_items():
d = {"a": 1, "b": 2, "c": 3}
for key in d:
val = d[key]
...
好的做法
def not_using_dict_items():
d = {"a": 1, "b": 2, "c": 3}
for key, val in d.items():
...
坏的做法
mytuple = 1, 2
x = mytuple[0]
y = mytuple[1]
好的做法
mytuple = 1, 2
x, y = mytuple
坏的做法
def timing_with_time():
start = time.time()
time.sleep(1)
end = time.time()
print(end - start)
好的做法是使用 time.perf_counter(),更精确:
def timing_with_time():
# more accurate
start = time.perf_counter()
time.sleep(1)
end = time.perf_counter()
print(end - start)
坏的做法
def print_vs_logging():
print("debug info")
print("just some info")
print("bad error")
好的做法
def print_vs_logging():
# versus
# in main
level = logging.DEBUG
fmt = '[%(levelname)s] %(asctime)s - %(message)s'
logging.basicConfig(level=level, format=fmt)
# wherever
logging.debug("debug info")
logging.info("just some info")
logging.error("uh oh :(")
坏的做法
subprocess.run(["ls -l"], capture_output=True, shell=True)
如果 shell=True,则将 ls -l
传递给/bin/sh(shell) 而不是 Unix 上的 ls 程序,会导致 subprocess 产生一个中间 shell 进程, 换句话说,使用中间 shell 意味着在命令运行之前,命令字符串中的变量、glob 模式和其他特殊的 shell 功能都会被预处理。比如,$HOME 会在在执行 echo 命令之前被处理处理。
好的做法是拒绝从 shell 执行:
subprocess.run(["ls", "-l"], capture_output=True)
坏的做法
def not_using_numpy_pandas():
x = list(range(100))
y = list(range(100))
s = [a + b for a, b in zip(x, y)]
好的做法:
import numpy as np
def not_using_numpy_pandas():
# 性能更快
x = np.arange(100)
y = np.arange(100)
s = x + y
坏的做法
from itertools import *
count()
这样的话,没有人直到这个脚本到底有多数变量, 好的做法:
from mypackage.nearby_module import awesome_function
def main():
awesome_function()
if __name__ == '__main__':
main()
欢迎转载、收藏、有所收获点赞支持一下!
目前开通了技术交流群,群友已超过2000人,添加时最好的备注方式为:来源+兴趣方向,方便找到志同道合的朋友