NumPy

Question 1

What is NumPy

Answer 1

A Python library for numerical data.

Question 2

What does NumPy provide functionality for?

Answer 2

Arrays, Linear Algebra, and RNG

Question 3

Why is important to use “import numpy as np” as opposed to “from numpy import *”?

Answer 3

Because NumPy namespace conflicts with python’s with functions like ‘min’ or ‘max’.

Question 4

How much faster are NumPy array operations vs standard python?

Answer 4

About 20 to 30 times faster.

Question 5

Why is NumPy so much faster than Python?

Answer 5

NumPy stores 1 object type and no need to dynamically create and reference variably sized memory addresses.

Question 6

How do you declare an array in numpy?

Answer 6

new_arr = np.array([x1,x2,…,xn])

Question 7

NumPy Boolean Operators: what does the following code produce?

a1 = np.array([[1., 2., 3.], [4., 5., 6.]])
a2 = np.array([[0., 4., 1.], [7., 2., 12.]])
a2 > a1

Answer 7

array([[False, True, False],
[ True, False, True]])

Question 8

Given nparray new_arr, what does new_arr.shape return?

Answer 8

A tuple containing the dimensions of new_arr.

Question 9

Given nparray new_arr, what does new_arr.dtype return?

Answer 9

The type of the data stored in new_arr.

Question 10

How do you specify the datatype in an nparray?

Answer 10

By using the overloaded constructor with the dtype flag. Ex: new_arr = np.array([1,2,3], dtype=np.int32)

Question 11

Will numpy allow creation of arrays with multiple data types?

Answer 11

Yes, but all indices will default to the common denominator (i.e. to a string or a float)

Question 12

What is default_rng?

Answer 12

A random number generator in numpy.

Question 13

How is default_rng imported?

Answer 13

from numpy.random import default_rng

Question 13

What does default_rng.integers(1,5, size=(10,)) produce?

Answer 13

A 1x10 array with int objects from 1 to 4.

Question 14

what does default_rng.random(size=(3,3)) produce?

Answer 14

A 3x3 array with float objects between 0.0 and 1.0.

Question 15

What does default_rng.standard_normal(size=(5,)) produce?

Answer 15

A 1x5 array with float objects that will produce a gaussian distribution over time.

Question 16

What does np.arange(1, 4, 0.1) produce?

Answer 16

A 1 dimensional array with all values of stride 0.1 between 1.0 and 3.9.

Question 17

What does np.linspace(1, 4, 2) produce?

Answer 17

A 1 dimensional array with all values of stride 0.2 between 1.0 and 4.0.

Question 18

Do list slices create a reference of a list or a copy?

Answer 18

A reference.

Question 19

Does fancy indexing create a reference of a list or a copy?

Answer 19

A copy.

Question 20

Given an array new_arr, how would you copy a slice of indices 5 through 8?

Answer 20

new__arr[5:9].copy()

Question 21

Given an array new_arr, how would you access the 2’nd row?

Answer 21

new_arr[1]

Question 22

Given an array new_arr, how would you access the second column?

Answer 22

new_arr[:, 1]

Question 23

Given an array new_arr, how would you extract the first three rows using fancy indexing?

Answer 23

new_arr[ [0,1,2] ]

Question 24

Given an array new_arr, how would you extract the first three columns using fancy indexing?

Answer 24

new_arr[ :, [0,1,2] ]

Question 24

What is Boolean indexing?

Answer 24

Selectively accessing parts of an array based upon a Boolean condition.

Question 25

What are the not, and, or operators in NumPy?

Answer 25

not = ~
and = &
or = |

Question 26

What does the reshape function do to a numpy array?

Answer 26

It modifies the shape of an array as long as no elements are lost or created.

Question 27

Given one dimensional numpy array num_arr with 15 elements, what does the following code produce?

num_arr.reshape((5, 3))

Answer 27

It takes the elements of num_arr and produces an array with 5 rows and 3 columns.

Question 28

What is the transpose of an array?

Answer 28

An array with all column and row indices flipped.

Question 29

What are universal functions (or ufuncs) in Numpy?

Answer 29

A function that performs element-wise operations on data in an np.array

Question 30

What is array broadcasting?

Answer 30

It is taking a vector and adding it to every row of an array.

Question 31

How would you find the sum of all elements in numpy array num_arr?

Answer 31

np.sum(num_arr) or num_arr.sum()

Question 32

How would you find the mean of all elements in numpy array num_arr?

Answer 32

np.mean(num_arr) or num_arr.mean()

Question 33

How would you find the standard deviation of all elements in numpy array num_arr?

Answer 33

np.std(num_arr) or num_arr.std()

Question 34

What is the difference between np.min and np.argmin or np.max and np.argmax?

Answer 34

min or max returns the value, argmin or argmax return the index of that value.

Question 35

What does np.sum(axis=0) return?

Answer 35

The sum of each column in an array.

Question 36

What does np.sum(axis=1) return?

Answer 36

The sum of each row in an array.

Question 37

Given num_arr = np.array([false, false, true, true, true]), what is returned by num_arr.sum()?

Answer 37

3