将数组作为输入传递给用户定义的函数

Question

If you want to pass vectors (or arrays) as input to some user defined function to make some operations with them, you can pass the components:

def sample(X_0,Y_0)
    #here X_0 and Y_0 are some components of the arrays X and Y
    #do some operations with those components of X and Y like
    scalar= X_0+Y_0
    return scalar
number=sample(X[0,0],Y[0,0])

or pass the vectors and decompose then inside the function:

def sample(X,Y)
    #here X and Y are arrays
    #do some operations like
    scalar=X[0,0]+Y[0,0]
    return scalar
number=sample(X,Y)

which is the preferred method and why?

Answer 1

In your specific case, there is absolutely no difference between the two outputs as you bind the results to scalar . In python, you tend to pass arrays and lists by reference and primitives by value. So when can you expect to see a difference?

Consider the function:

>>> def manipulateX(X):
...     X[len(X)-1] = -1
...     return X
... 

When you call manipulateX with a list (as follows), you see that the list is also manipulated:

>>> x = [1, 2, 3, 4]
>>> manipulateX(x)
[1, 2, 3, -1]
>>> x
[1, 2, 3, -1]

However, if you define a function which operates on a primitive

>>> def manipulateY(Y):
...     Y += 20
...     return Y
... 

and call it with an item in a collection (list, array) etc:

>>> manipulateY(x[0])
21
>>> x
[1, 2, 3, -1]
>>> 

You see that x[0] remains unchanged. In your case as you bind your results to scalar you don't see any difference. In both the cases, your memory usage is also the same. If you weren't binding the results to scalar then it depends on whether you want x[0] and y[0] to be muted.