Map Function for Numpy Matrix

Question

I have two numpy matrices. One contains lambda functions. The other contains values.

Is there a function that is similar to Python's map function that will allow me to get the expected result?

Is there a better way?

functionMatrix = np.array([[lambda x:x**2, lambda x:x**3],[lambda x: x**2, 
lambda x: np.sqrt(x)]])
valueMatrix = np.array([[1,2],[3,4]])

expectedResult = np.array([[1,8],[9,2]])

If it's always simple exponents you're trying to get, you could just avoid the lambdas, and do something like : functionMatrix = np.array([[2, 3],[2, 0.5]]) and then valueMatrix ** functionMatrix (output: array([[ 1., 8.], [ 9., 2.]])). Also, judging by your expected output, I think you mean lambda x: x**2 instead of lambda x: x*2, and np.sqrt instead of np.square — sacuL
– sacuL, Commented May 8, 2018 at 20:27
sacul - You're right, let me fix that. It won't always be simple exponents. Some functions might be natural log, inverse, etc. — David
– David, Commented May 8, 2018 at 20:33
"is there a better way?" depends on if there's a mathematical relationship to all your lambda functions. Putting them inside a numpy array nets you an 'object array' which is a convenience feature built into numpy, but it does not take advantage of many of the benefits of a normal array (from a speed or memory standpoint). — Aaron
– Aaron, Commented May 8, 2018 at 20:47

hilberts_drinking_problem · Accepted Answer · 2018-05-08 20:36:10Z

3

This is just syntactic sugar but does the job.

@np.vectorize
def apply_vec(f, x):
    return f(x)

result = apply_vec(functionMatrix, valueMatrix)

answered May 8, 2018 at 20:36

hilberts_drinking_problem

11.6k3 gold badges25 silver badges55 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Map Function for Numpy Matrix

1 Answer 1

Comments

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Related