The error “TypeError: expected sequence, got ndarray” occurs when a function expects a list, tuple, or other sequence-like data structure but receives a NumPy ndarray.
Possible Causes and Fixes:
1. Using NumPy Array Instead of List
Some functions in Python, especially those expecting iterable sequences, may not work directly with NumPy arrays.
Incorrect:
import numpy as np
arr = np.array([1, 2, 3])
print(sum(arr)) # TypeError
Fix: Convert the ndarray to a list before using the function.
print(sum(arr.tolist()))
2. Passing NumPy Array to a Function Expecting a Sequence
Certain machine learning libraries, such as Scikit-learn, expect a sequence (list) rather than an ndarray.
Incorrect:
from sklearn.preprocessing import MinMaxScaler
import numpy as np
scaler = MinMaxScaler()
X = np.array([1, 2, 3, 4, 5])
X_scaled = scaler.fit_transform(X) # TypeError
Fix: Reshape the array to 2D.
X_scaled = scaler.fit_transform(X.reshape(-1, 1))
3. Providing a NumPy Array Instead of a List to a Pandas DataFrame
Pandas expects sequences like lists when creating a DataFrame.
Incorrect:
import pandas as pd
data = np.array([1, 2, 3])
df = pd.DataFrame(data, columns="A") # TypeError
Fix:
df = pd.DataFrame(data, columns=["A"])
4. Using NumPy Array Instead of List in zip()
The zip() function expects iterable sequences, not ndarray.
Incorrect:
arr1 = np.array([1, 2, 3])
arr2 = np.array([4, 5, 6])
zipped = zip(arr1, arr2) # TypeError
Fix: Convert to lists.
zipped = zip(arr1.tolist(), arr2.tolist())