Type casting with astype() in NumPy

Introduction

Type casting changes the data type of array elements. It helps when you need numbers as integers or floats for calculations.

You have decimal numbers but want to work with whole numbers only.

You want to save memory by converting large float arrays to smaller integer types.

You need to convert numbers to strings for display or export.

You want to prepare data for a function that requires a specific type.

You want to convert boolean values to integers for counting.

Syntax

NumPy

array.astype(new_type)

array is your NumPy array.

new_type can be a type like int, float, str, or NumPy types like np.int32.

Examples

Converts float numbers to integers by dropping decimals.

NumPy

import numpy as np
arr = np.array([1.5, 2.3, 3.7])
arr_int = arr.astype(int)
print(arr_int)

Converts integers to boolean values (0 becomes False, others True).

NumPy

import numpy as np
arr = np.array([1, 0, 1])
arr_bool = arr.astype(bool)
print(arr_bool)

Converts numbers to strings.

NumPy

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

Sample Program

This program shows how to change the type of a NumPy array from float to int, string, and boolean using astype().

NumPy

import numpy as np

# Original array with floats
arr = np.array([4.7, 5.2, 6.9])
print('Original array:', arr)

# Convert to integers
arr_int = arr.astype(int)
print('As integers:', arr_int)

# Convert to strings
arr_str = arr.astype(str)
print('As strings:', arr_str)

# Convert to boolean
arr_bool = arr_int.astype(bool)
print('As booleans:', arr_bool)

Important Notes

When converting floats to integers, decimals are dropped (not rounded).

Converting to boolean treats zero as False and any other number as True.

Use astype() to create a new array; the original array stays the same.

Summary

astype() changes the data type of a NumPy array.

It is useful to prepare data for different tasks or save memory.

Always remember it returns a new array and does not change the original.

Practice

1. What does the astype() method do in NumPy?

easy

A. It changes the data type of a NumPy array and returns a new array.

B. It sorts the elements of a NumPy array in ascending order.

C. It removes duplicate elements from a NumPy array.

D. It reshapes the NumPy array into a new shape.

Solution

1. Step 1: Understand the purpose of astype()

   The astype() method is used to convert the data type of elements in a NumPy array.

2. Step 2: Recall the behavior of astype()

   It returns a new array with the changed data type and does not modify the original array.

3. Final Answer:

   It changes the data type of a NumPy array and returns a new array. -> Option A

4. Quick Check:

   astype() changes type and returns new array [OK]

Hint: Remember: astype() changes type and returns a new array [OK]

Common Mistakes:

• Thinking astype() modifies the original array in place
• Confusing astype() with sorting or reshaping functions
• Assuming astype() removes duplicates

2. Which of the following is the correct syntax to convert a NumPy array arr to float type using astype()?

easy

A. arr.convert(float)

B. astype(arr, float)

C. arr.to_float()

D. arr.astype(float)

Solution

1. Step 1: Recall the method call syntax for astype()

   The astype() method is called on the array object like arr.astype().

2. Step 2: Check the argument for data type

   The argument should be the desired data type, here float, passed inside parentheses.

3. Final Answer:

   arr.astype(float) -> Option D

4. Quick Check:

   Correct syntax is arr.astype(type) [OK]

Hint: Call astype() on array with desired type inside parentheses [OK]

Common Mistakes:

• Using astype as a standalone function instead of a method
• Using incorrect method names like convert or to_float
• Passing data type outside parentheses

3. What will be the output of the following code?

import numpy as np
arr = np.array([1.7, 2.3, 3.9])
new_arr = arr.astype(int)
print(new_arr)

medium

A. [2 3 4]

B. [1.7 2.3 3.9]

C. [1 2 3]

D. Error: Cannot convert float to int

Solution

1. Step 1: Understand astype(int) on float array

   Converting floats to int using astype(int) truncates the decimal part (does not round).

2. Step 2: Apply truncation to each element

   1.7 becomes 1, 2.3 becomes 2, 3.9 becomes 3.

3. Final Answer:

   [1 2 3] -> Option C

4. Quick Check:

   astype(int) truncates floats to integers [OK]

Hint: astype(int) truncates decimals, does not round [OK]

Common Mistakes:

• Assuming astype(int) rounds instead of truncates
• Expecting original array to change
• Thinking conversion causes error

4. Identify the error in the following code snippet:

import numpy as np
arr = np.array([1, 2, 3])
new_arr = arr.astype('int32')
print(new_arr.astype('float64'))

medium

A. Error because 'int32' is not a valid dtype string

B. No error; code runs and prints float64 array

C. Error because astype() cannot be chained

D. Error because original array is modified

Solution

1. Step 1: Check dtype string validity

   'int32' and 'float64' are valid NumPy dtype strings for 32-bit int and 64-bit float.

2. Step 2: Verify chaining of astype()

   Chaining astype() calls is allowed because each returns a new array.

3. Final Answer:

   No error; code runs and prints float64 array -> Option B

4. Quick Check:

   Valid dtype strings and chaining astype() work [OK]

Hint: astype() returns new array; chaining is allowed [OK]

Common Mistakes:

• Thinking dtype strings like 'int32' are invalid
• Believing astype() modifies original array
• Assuming astype() cannot be chained

5. You have a NumPy array arr = np.array([10, 20, 30]) of integers. You want to save memory by converting it to the smallest unsigned integer type that can hold these values. Which astype() call is best?

hard

A. arr.astype('uint8')

B. arr.astype('int8')

C. arr.astype('float16')

D. arr.astype('uint16')

Solution

1. Step 1: Check the range of values in arr

   Values are 10, 20, 30, all positive and less than 255.

2. Step 2: Choose smallest unsigned integer type

   uint8 holds values 0 to 255, so it fits all values and saves memory.

3. Step 3: Compare with other options

   int8 supports -128 to 127 but is signed; float16 uses more memory; uint16 is larger than needed.

4. Final Answer:

   arr.astype('uint8') -> Option A

5. Quick Check:

   Use smallest unsigned int type fitting data [OK]

Hint: Pick smallest unsigned int type covering max value [OK]

Common Mistakes:

• Choosing signed int type that wastes range or causes errors
• Using float type which uses more memory
• Ignoring value ranges when selecting dtype