In this article, we will see how to fix TypeError: Object of type ndarray is not JSON serializable when we are converting a NumPy ndarray object to a JSON string.
What is “TypeError: Object of type ndarray is not JSON serializable”?
The error “TypeError: Object of type ndarray is not JSON serializable” generally occurs in Python when we are converting a NumPy ndarray object to a JSON String. In this example, the demonstration of this error is shown below.
Python3
import jsonimport numpy as nparr = np.array([1, 2, 3])#TypeError: Object of type ndarray is not JSON serializablejson_str = json.dumps({'nums': arr}) |
Output
TypeError: Object of type ndarray is not JSON serializable
Type Error: NumPy Array is not JSON Serializable
Below are the ways by which we can fix this error and handle the conversion of ndarray object to a JSON string.
- Using tolist()
- Extending the JSONEncoder class
- Using the default keyword argument
- Using pandas library
Using tolist() to fix TypeError
A crucial function provided by Python NumPy, which is used to convert array to list is known as tolist() function. In this way, we will see how we can fix the Type Error: NumPy array is not JSON serializable by converting NumPy array to list. In this example, we are using tolist() to convert NumPy Array to list.
Python3
# Python program to Fix Type Error : NumPy array is not # JSON serializable converting NumPy array to listimport jsonimport numpy as np# Create a NumPy arrayarray = np.array([7, 4, 2])# Serialize JSON after converting NumPy array to listarr_json = json.dumps({'nums': array.tolist()})# Print the serialized JSONprint(arr_json) |
Output
Extending the JSONEncoder class
The other way of fixing the type error: NumPy array is not JSON serializable, is by creating a function inside a class, and call that class while serializing the JSON. This function will convert NumPy array to JSON. In this example, we are extending the JSONEncoder class to fix this error.
Python3
# Import the libraries json and numpyimport jsonimport numpy as np# Create a JSON Encoder classclass json_serialize(json.JSONEncoder): def default(self, obj): if isinstance(obj, np.integer): return int(obj) if isinstance(obj, np.floating): return float(obj) if isinstance(obj, np.ndarray): return obj.tolist() return json.JSONEncoder.default(self, obj)# Create a NumPy arrayarray = np.array([7, 4, 2])# Serialize JSON by calling JSON Encoder classjson_array = json.dumps({'nums': array}, cls=json_serialize)# Print the serialized JSONprint(json_array) |
Output
Using the default keyword argument
The json.dumps() offers various parameters out of which default is the significant parameter which helps in JSON serializing the NumPy array. In this way, we will see how can we JSON serialize NumPy array and fix Type Error: NumPy array is not JSON serializable by using default keyword argument. In this example, we are using default keyword to solve this error.
Python3
# Import the libraries json and numpyimport jsonimport numpy as np# Create a function to be called while serializing JSONdef json_serialize(obj): if isinstance(obj, np.ndarray): return obj.tolist() return obj# Create a NumPy arrayarray = np.array([7, 4, 2])# Serialize JSON by using default keywordjson_array = json.dumps({'nums': array}, default=json_serialize)# Print the serialized JSONprint(json_array) |
Output
Using pandas library
Pandas is the library which is commonly used for working with datasets. A crucial function offered by Pandas is Series(), which is like a 1D array and can hold data of any type. In this way, we will see how can we JSON serialize NumPy array and fix Type Error: NumPy array is not JSON serializable by using Pandas Series function. In this example, we are using pandas library to solve this error.
Python3
# Import the libraries NumPy and Pandasimport numpy as npimport pandas as pd# Create a NumPy arrayarray = np.array([7, 4, 2])# Serialize JSON by using Pandasjson_array = pd.Series(array).to_json(orient='values')# Print the serialized JSONprint('{"nums": ' +json_array+'}') |
Output
