Skip to main content

Python Numpy: How to Fix AttributeError: 'list' object has no attribute 'shape' / 'reshape' / 'astype'

When working with numerical data in Python, especially when transitioning from standard Python lists to NumPy arrays for more advanced computations, you might encounter AttributeError: 'list' object has no attribute 'shape', 'reshape', or 'astype'. This error is a clear indication that you are attempting to use a NumPy-specific attribute or method on a standard Python list object, which does not inherently possess these functionalities.

This guide will thoroughly explain why this AttributeError occurs, demonstrate common scenarios where these NumPy operations are mistakenly applied to lists, and provide the straightforward solution: convert your Python list to a NumPy array using np.array() or np.asarray() before attempting to use these array-specific features.

Understanding the Error: Python Lists vs. NumPy Arrays​

  • Python list: A built-in, general-purpose, ordered collection of items. Lists are very flexible but do not have inherent concepts like multi-dimensional shape, reshaping methods, or bulk data type conversion methods common in numerical libraries.
  • NumPy ndarray (array): A fundamental object in the NumPy library designed for efficient numerical computation. NumPy arrays have a shape attribute (describing their dimensions), a reshape() method (to change dimensions), and an astype() method (to change the data type of all elements).

The AttributeError: 'list' object has no attribute 'X' occurs because you are trying to call method X or access attribute X (which belongs to NumPy arrays) directly on a Python list variable.

AttributeError: 'list' object has no attribute 'shape'​

This means you're trying to get the dimensions of a list using the NumPy array syntax .shape.

Reproducing the Error​

my_python_list = [
    [10, 20, 30],
    [40, 50, 60]
]

try:
    # ⛔️ Incorrect: Python lists do not have a .shape attribute
    list_shape = my_python_list.shape
    print(list_shape)
except AttributeError as e:
    print(f"Error: {e}")

Output:

Error: 'list' object has no attribute 'shape'

Solution: Convert List to NumPy Array to Get shape​

First, convert the list to a NumPy array.

import numpy as np

my_python_list = [[10, 20, 30], [40, 50, 60]]

# ✅ Convert list to NumPy array
numpy_array_from_list = np.array(my_python_list)

# Now .shape attribute is available
array_shape = numpy_array_from_list.shape
print(f"Shape of the NumPy array: {array_shape}")

Output:

Shape of the NumPy array: (2, 3)
note

You can also use np.asarray() which avoids a copy if the input is already an ndarray:

numpy_array_asarray = np.asarray(my_python_list)
print(f"Shape using asarray: {numpy_array_asarray.shape}")

Alternative: Using numpy.shape() Directly on a List​

The numpy.shape() function can take a list-like object as input and return its shape as if it were a NumPy array.

import numpy as np

my_python_list = [[10, 20, 30], [40, 50, 60]]

# ✅ Pass the list directly to np.shape()
shape_via_np_shape = np.shape(my_python_list)
print(f"Shape of list via np.shape(): {shape_via_np_shape}")

print(f"Shape of [1,2,3,4] via np.shape(): {np.shape([1,2,3,4])}")

Output:

Shape of list via np.shape(): (2, 3)
Shape of [1,2,3,4] via np.shape(): (4,)

For Python Lists: Using len() for Length​

If you simply want the number of elements in a list (or number of rows for a list of lists), use the built-in len() function.

my_python_list = [[10, 20, 30], [40, 50, 60]]

num_rows = len(my_python_list)
print(f"Number of rows (outer list length): {num_rows}")

if num_rows > 0:
    num_cols_in_first_row = len(my_python_list[0])
    print(f"Number of columns in the first row: {num_cols_in_first_row}")

Output:

Number of rows (outer list length): 2
Number of columns in the first row: 3
note

This doesn't give a "shape" tuple like NumPy but provides dimensional information.

AttributeError: 'list' object has no attribute 'reshape'​

This error means you're trying to use NumPy's reshape() method on a Python list.

Reproducing the Error​

flat_list = [1, 2, 3, 4, 5, 6, 7, 8]

try:
    # ⛔️ Incorrect: Python lists do not have a .reshape() method
    reshaped_list_error = flat_list.reshape((2, 4))
    print(reshaped_list_error)
except AttributeError as e:
    print(f"Error: {e}")

Output:

Error: 'list' object has no attribute 'reshape'

Solution: Convert List to NumPy Array Before Calling reshape()​

import numpy as np

flat_list = [1, 2, 3, 4, 5, 6, 7, 8]

# ✅ Convert list to NumPy array first
numpy_array_for_reshape = np.array(flat_list)

# Now .reshape() method is available
reshaped_array = numpy_array_for_reshape.reshape((2, 4))
print("Reshaped NumPy array (2x4):")
print(reshaped_array)

Output:

Reshaped NumPy array (2x4):
[[1 2 3 4]
 [5 6 7 8]]

Alternative: Using numpy.reshape() Directly on a List​

The numpy.reshape() function can take a list-like object as its first argument.

import numpy as np

flat_list = [1, 2, 3, 4, 5, 6, 7, 8]

# ✅ Pass the list directly to np.reshape()
array_reshaped_direct = np.reshape(flat_list, (4, 2)) # Reshape to 4x2
print("Array reshaped directly by np.reshape (4x2):")
print(array_reshaped_direct)

# Reshaping a list of lists (already 2D-like)
list_of_lists = [[1,2,3],[4,5,6]] # 2x3
reshaped_from_lol = np.reshape(list_of_lists, (3,2)) # Reshape to 3x2
print("List of lists reshaped to 3x2:")
print(reshaped_from_lol)

Output:

Array reshaped directly by np.reshape (4x2):
[[1 2]
 [3 4]
 [5 6]
 [7 8]]
List of lists reshaped to 3x2:
[[1 2]
 [3 4]
 [5 6]]

AttributeError: 'list' object has no attribute 'astype'​

This error indicates an attempt to use NumPy's astype() method (for changing data types) on a Python list.

Reproducing the Error​

mixed_type_list = [10, 20.5, '30', 40.0]

try:
    # ⛔️ Incorrect: Python lists do not have an .astype() method
    list_astype_error = mixed_type_list.astype(int)
    print(list_astype_error)
except AttributeError as e:
    print(f"Error: {e}")

Output:

Error: 'list' object has no attribute 'astype'

Solution: Convert List to NumPy Array Before Calling astype()​

import numpy as np

mixed_type_list = [10, 20.5, '30', 40.0, '55.5'] # '30' and '55.5' are strings

# ✅ Convert list to NumPy array first
# When creating, NumPy might infer a common dtype (e.g., object or string if mixed)
numpy_array_for_astype = np.array(mixed_type_list)
print(f"NumPy array from mixed list (initial dtype: {numpy_array_for_astype.dtype}):")
print(numpy_array_for_astype)
print()

# Now .astype() method is available to change type (e.g., to float, then to int)
array_as_float = numpy_array_for_astype.astype(float)
print(f"Array after astype(float): {array_as_float}")

array_as_int = array_as_float.astype(int) # Floats will be truncated
print(f"Array after astype(int): {array_as_int}")

Output:

NumPy array from mixed list (initial dtype: <U32):
['10' '20.5' '30' '40.0' '55.5']

Array after astype(float): [10.  20.5 30.  40.  55.5]
Array after astype(int): [10 20 30 40 55]
note

Alternatively, specify dtype during np.array creation if a direct conversion is possible. For mixed_type_list, direct to int would fail for '20.5'.

import numpy as np

list_of_numbers_as_strings = ['11', '22', '33']
array_direct_dtype = np.array(list_of_numbers_as_strings, dtype=int)
print(f"Array created with direct dtype=int: {array_direct_dtype}")

Output:

Array created with direct dtype=int: [11 22 33]

Debugging Tip: Checking Object Attributes with dir()**​

If you're unsure what attributes or methods an object (like your list or array variable) has, you can use the built-in dir() function.

my_simple_list = [1, 'two', 3.0]
print("Attributes of a Python list:")
print([attr for attr in dir(my_simple_list) if not attr.startswith('__')]) # Filter out dunder methods

import numpy as np
my_numpy_array = np.array([1,2,3])
print("Attributes of a NumPy array (partial list):")
print([attr for attr in dir(my_numpy_array) if not attr.startswith('__') and attr in ['shape', 'reshape', 'astype', 'T', 'sum']])

Output:

Attributes of a Python list:
['append', 'clear', 'copy', 'count', 'extend', 'index', 'insert', 'pop', 'remove', 'reverse', 'sort']
Attributes of a NumPy array (partial list):
['T', 'astype', 'reshape', 'shape', 'sum']
note

Comparing the outputs of dir() for a list and a NumPy array clearly shows why these AttributeErrors occur.

Conclusion​

The AttributeError: 'list' object has no attribute 'shape', 'reshape', or 'astype' is a common signal that you are trying to apply NumPy array-specific functionalities to standard Python lists. The universal solution is to first convert your Python list into a NumPy array using numpy.array(your_list) or numpy.asarray(your_list).

  • Once you have a NumPy array object, attributes like .shape and methods like .reshape() and .astype() will be available and work as expected.
  • For some functions like np.shape() and np.reshape(), NumPy can directly accept list-like input, providing a more concise alternative to explicit conversion first.

Table of Contents