Creating sparse matrices in SciPy - Practice Exercises

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of COO sparse matrix data arrays

What will be the output of the following code snippet that creates a COO sparse matrix and prints its data, row, and column arrays?

SciPy

from scipy.sparse import coo_matrix
import numpy as np

row = np.array([0, 3, 1, 0])
col = np.array([0, 3, 1, 2])
data = np.array([4, 5, 7, 9])

matrix = coo_matrix((data, (row, col)), shape=(4, 4))
print(matrix.data)
print(matrix.row)
print(matrix.col)

A

[4 9 7 5]
[0 0 1 3]
[0 2 1 3]

B

[9 7 5 4]
[3 1 0 0]
[2 1 3 0]

C

[4 5 7 9]
[0 3 1 0]
[0 3 1 2]

D

[4 5 7 9]
[0 1 2 3]
[0 1 2 3]

💡 Hint

Remember that COO matrix stores data in the order you provide the row and column indices.

The COO sparse matrix stores the data, row, and column arrays exactly as given. So the printed arrays match the input arrays without reordering.

❓ data_output

intermediate

Number of non-zero elements in CSR matrix

Given the following CSR sparse matrix creation code, how many non-zero elements does the matrix contain?

SciPy

from scipy.sparse import csr_matrix

values = [10, 20, 30, 40]
col_indices = [0, 2, 2, 0]
row_ptr = [0, 2, 3, 4]

csr = csr_matrix((values, col_indices, row_ptr), shape=(3, 3))
print(csr.nnz)

A 4

B 3

C 2

D 5

💡 Hint

The nnz attribute gives the count of stored non-zero elements.

The nnz attribute of a CSR matrix returns the number of stored non-zero elements, which is 4 here.

❓ visualization

advanced

Visualizing a sparse matrix pattern

Which option shows the correct matplotlib visualization code to display the sparsity pattern of a CSR matrix named 'csr'?

SciPy

import matplotlib.pyplot as plt
from scipy.sparse import csr_matrix

csr = csr_matrix([[0, 0, 1], [1, 0, 0], [0, 2, 0]])

A

plt.imshow(csr)
plt.show()

B

plt.spy(csr)
plt.show()

C

plt.plot(csr)
plt.show()

D

plt.scatter(csr)
plt.show()

💡 Hint

Use the function designed to visualize sparse matrix structure.

plt.spy is the correct function to visualize the sparsity pattern of a sparse matrix.

🧠 Conceptual

advanced

Memory efficiency of sparse matrix formats

Which sparse matrix format is generally most memory efficient for matrices with many rows but very few non-zero elements per row?

A CSR (Compressed Sparse Row)

B CSC (Compressed Sparse Column)

C COO (Coordinate format)

D DOK (Dictionary of Keys)

💡 Hint

Think about fast row slicing and compact storage of row pointers.

CSR format stores row pointers compactly and is efficient for matrices with many rows and few non-zero elements per row.

🔧 Debug

expert

Identify the error in sparse matrix creation

What error will this code raise when trying to create a CSR matrix?

SciPy

from scipy.sparse import csr_matrix

values = [1, 2, 3]
col_indices = [0, 1]
row_ptr = [0, 2, 3]

csr = csr_matrix((values, col_indices, row_ptr), shape=(2, 3))

A IndexError: index out of bounds

B No error, matrix created successfully

C TypeError: unsupported operand type(s)

D ValueError: array length mismatch

💡 Hint

Check if the lengths of values and column indices match.

The values array length (3) does not match the column indices length (2), causing a ValueError.