Python Performance Optimization Tips

Mastering a few techniques can significantly improve Python program performance and avoid unnecessary resource waste.

1. Use local variables – Prefer local over global variables to enhance maintainability, speed, and memory usage. For example, assign ls = os.linesep instead of repeatedly accessing os.linesep.

2. Reduce function call count – When checking an object's type, isinstance() is the most efficient, followed by identity comparison with is, and finally type(). Example:

#判断变量num是否为整数类型
type(num) == type(0)  # calls three functions
type(num) is type(0)   # identity comparison
isinstance(num, (int)) # calls one function

3. Avoid repeated operations in loop conditions – Compute expensive expressions like len(a) once before the loop:

#每次循环都需要重新执行len(a)
while i < len(a):
    statement
#len(a)仅执行一次
m = len(a)
while i < m:
    statement

4. Import specific objects directly – Use from X import Y instead of import X; X.Y to save one lookup.

5. Use mappings instead of conditional chains – Dictionary lookups are far faster than multiple if/elif statements. Example:

#dict查找，性能更优
d = {1:10, 2:20, ...}
b = d[a]

6. Iterate over sequence elements directly – Looping over items is quicker than looping over indices:

a = [1,2,3]
for item in a:
    print(item)
# versus
for i in range(len(a)):
    print(a[i])

7. Prefer generator expressions over list comprehensions – Generators produce items lazily, reducing memory usage for large data sets:

#生成器表达式
l = sum([len(word) for line in f for word in line.split()])
#列表解析（会创建完整列表）
l = sum(len(word) for line in f for word in line.split())

8. Compile code before execution – Use compile() to obtain a code object and then call eval() or exec(), avoiding repeated compilation. The same applies to regular expressions with re.compile().

9. Adopt module programming habits – Place executable code inside functions and guard test code with if __name__ == '__main__': to prevent unwanted execution on import.

The article concludes with a QR code offering a free Python public course and additional learning resources.