lambdify integral solution to inverse power law

Question

I am using sympy to evaluate an integral of the form 1/t^p and then lambdify the expression to evaluate it. Evaluated at p=1, the resolution expression should return ln(t), but instead returns a ZeroDivisionError. I expect I will need to expand the solution around p=1. Is that necessary and how does one go about this?

>> expr = integrate(1/t**p,t)
>> expr
Out: 
⎧ 1 - p           
⎪t                
⎪──────  for p ≠ 1
⎨1 - p            
⎪                 
⎪log(t)  otherwise
⎩                 

>> lam_expr = lambdify((t,p),expr)
>> lam_expr(10,1)
Traceback (most recent call last):

  File "<ipython-input-84-8763c416d5cb>", line 1, in <module>
    lam_expr(10,1)

  File "<lambdifygenerated-33>", line 2, in _lambdifygenerated
    return (select([not_equal(p, 1),True], [t**(1 - p)/(1 - p),log(t)], default=nan))

ZeroDivisionError: division by zero

Update

When using a NumPy array I do not encounter the error though the warning is still issued

>> In: lam_expr(np.array([1,2,3,5]),1)
<lambdifygenerated-101>:2: RuntimeWarning: divide by zero encountered in true_divide
  return (select([not_equal(p, 1),True], [t**(1 - p)/(1 - p),log(t)], default=nan))
>> Out: array([0.        , 0.69314718, 1.09861229, 1.60943791])

It seems like a (bad) solution to this problem would be to turn all input into NumPy arrays.