Why Array attributes (shape, dtype, ndim, size) in NumPy? - Purpose & Use Cases

The Big Idea

Discover how a few simple attributes can save you hours of manual counting and guessing in your data work!

The Scenario

Imagine you have a big table of numbers written on paper. You want to know how many rows and columns it has, what kind of numbers are inside, and how many numbers there are in total. Doing this by counting each row and column manually is tiring and slow.

The Problem

Counting rows and columns by hand is easy to mess up, especially if the table is huge or has many layers. Also, figuring out if the numbers are whole numbers or decimals by looking is error-prone. This wastes time and can lead to wrong answers.

The Solution

Using array attributes like shape, dtype, ndim, and size in NumPy instantly tells you the structure and type of your data. This saves time, avoids mistakes, and helps you understand your data quickly.

Before vs After

✗ Before

rows = len(data)
cols = len(data[0])
type_of_data = type(data[0][0])
total_elements = rows * cols

✓ After

rows, cols = array.shape
data_type = array.dtype
dimensions = array.ndim
total_elements = array.size

What It Enables

With these attributes, you can quickly explore and prepare your data for analysis or machine learning without guessing or manual counting.

Real Life Example

Suppose you receive a dataset of images stored as arrays. Using shape tells you the image size, dtype tells you if pixels are stored as integers or floats, ndim tells you if images are grayscale or color, and size tells you the total pixels per image.

Key Takeaways

shape tells the size of each dimension of the array.

dtype tells the type of data stored (like integer or float).

ndim tells how many dimensions the array has.

size tells the total number of elements in the array.

Practice

1. Which numpy attribute tells you the total number of elements in an array?

easy

A. size

B. shape

C. ndim

D. dtype

Solution

1. Step 1: Understand the meaning of each attribute

   shape shows dimensions, dtype shows data type, ndim shows number of dimensions, and size shows total elements.

2. Step 2: Identify the attribute for total elements

   The attribute size gives the total number of elements in the array.

3. Final Answer:

   size -> Option A

4. Quick Check:

   Total elements = size [OK]

Hint: Total elements? Use size attribute [OK]

Common Mistakes:

• Confusing shape with total elements
• Thinking ndim gives total elements
• Mixing dtype with size

2. Which of the following is the correct way to get the number of dimensions of a numpy array arr?

easy

A. arr.size()

B. arr.ndim

C. arr.shape()

D. arr.dtype

Solution

1. Step 1: Recall attribute vs method syntax in numpy

   Attributes like ndim and dtype are accessed without parentheses, methods like size() do not exist.

2. Step 2: Identify correct attribute for number of dimensions

   arr.ndim gives the number of dimensions directly.

3. Final Answer:

   arr.ndim -> Option B

4. Quick Check:

   Number of dimensions = arr.ndim [OK]

Hint: Use attribute without () for ndim, size, dtype [OK]

Common Mistakes:

• Using parentheses with attributes
• Confusing shape with ndim
• Trying to call size() as method

3. Given the code:

import numpy as np
arr = np.array([[1, 2, 3], [4, 5, 6]])
print(arr.shape, arr.dtype, arr.ndim, arr.size)

What is the output?

medium

A. (2, 3) float64 2 6

B. (3, 2) int64 2 6

C. (2, 3) int64 2 6

D. (2, 3) int32 2 6

Solution

1. Step 1: Analyze the array shape and data type

   The array has 2 rows and 3 columns, so shape is (2, 3). The default integer type on many systems is int64.

2. Step 2: Check number of dimensions and total size

   It is a 2D array, so ndim = 2. Total elements = 2*3 = 6.

3. Final Answer:

   (2, 3) int64 2 6 -> Option C

4. Quick Check:

   shape=(2,3), dtype=int64, ndim=2, size=6 [OK]

Hint: Shape = rows x columns, size = product of shape [OK]

Common Mistakes:

• Mixing rows and columns in shape
• Assuming dtype is int32 by default
• Confusing ndim with size

4. What is wrong with this code snippet?

import numpy as np
arr = np.array([1, 2, 3])
print(arr.shape())

medium

A. print cannot display tuples

B. arr is not defined

C. shape returns a number, not a tuple

D. shape is an attribute, not a method; remove parentheses

Solution

1. Step 1: Identify usage of shape

   shape is an attribute, so it should be accessed without parentheses.

2. Step 2: Correct the code

   Change arr.shape() to arr.shape to fix the error.

3. Final Answer:

   shape is an attribute, not a method; remove parentheses -> Option D

4. Quick Check:

   Attributes have no () [OK]

Hint: Attributes like shape have no () [OK]

Common Mistakes:

• Using parentheses with attributes
• Thinking shape returns a single number
• Assuming print can't show tuples

5. You have a numpy array arr with shape (4, 5, 6). Which attribute would you use to find the total number of elements, and what is that number?

hard

A. arr.size, 120

B. arr.ndim, 3

C. arr.shape, (4, 5, 6)

D. arr.dtype, depends on data

Solution

1. Step 1: Understand the attributes

   size gives total elements, ndim gives number of dimensions, shape gives dimensions, dtype gives data type.

2. Step 2: Calculate total elements from shape

   Total elements = 4 * 5 * 6 = 120, so arr.size = 120.

3. Final Answer:

   arr.size, 120 -> Option A

4. Quick Check:

   Total elements = product of shape = 120 [OK]

Hint: Multiply shape dimensions or use size [OK]

Common Mistakes:

• Confusing ndim with total elements
• Using shape instead of size
• Ignoring that dtype is unrelated to size