Sympy无法化简正弦欧拉公式？为何结果不为零？余弦可正常化简

Great question! Let's break down why this happens and how to fix it.

First, the core issue here is that SymPy's default simplify() function uses a general-purpose set of rules, and it doesn't always prioritize converting complex exponential expressions with 1/I terms into into trigonometric functions. Here's the breakdown:

Why Cosine Works But Sine Doesn't #

The Euler formula for cosine is (e^(ix) + e^(-ix))/2—this expression doesn't involve a reciprocal of the imaginary unit I, so SymPy's default simplify logic easily recognizes it as cos(x).

For sine, though, the formula is (e^(ix) - e^(-ix))/(2i), which is equivalent to -i*(e^(ix)-e^(-ix))/2. The 1/I term here is a small but crucial detail that the default simplify() doesn't automatically handle in this context, so it fails to reduce expr - sin(x) to zero.

Fixes to Get the Expected Zero Result #

You have a few straightforward ways to make SymPy recognize the equivalence:

1. Use trigsimp() (Trigonometry-Specific Simplification) #

This function is designed explicitly to handle conversions between trigonometric functions and complex exponentials. Replace simplify() with trigsimp():

from sympy import *
x = Symbol("x")
expr = (1/(2*I)) * (E**(I*x) - E**(-I*x))
print(trigsimp(expr - sin(x)))  # Output: 0

2. Expand the Complex Expression with expand_complex() #

This function expands complex terms and simplifies them, which will resolve the 1/I term and match it to sin(x):

print(expand_complex(expr - sin(x)))  # Output: 0

3. Manually Simplify the 1/(2*I) Term First #

Since 1/I = -I, you can rewrite your expression to avoid the reciprocal of I upfront. SymPy's default simplify() will then work as expected:

expr_simplified = (-I/2) * (E**(I*x) - E**(-I*x))
print(simplify(expr_simplified - sin(x)))  # Output: 0

The key takeaway is that while simplify() is great for general cases, using specialized functions like trigsimp() or expand_complex() gives you better control when working with trigonometric and complex exponential conversions.

内容的提问来源于stack exchange，提问作者user9730851