Python is a popular programming language used for various applications. It provides developers with an extensive library of modules and functions to help them achieve their programming goals. However, when working with data, you may encounter the error “Object of type ndarray is not JSON serializable.” In this article, we will explore what this error means and how to fix it.
Understanding the Error
The error message “Object of type ndarray is not JSON serializable” indicates that the data you are trying to serialize into JSON format contains an object of type ndarray, which is not supported by the JSON encoder. Ndarray stands for n-dimensional array, which is a type of object used to store and manipulate data in Python’s NumPy library. JSON, on the other hand, is a lightweight data interchange format that is commonly used to transmit data over the internet.
When you try to serialize an ndarray object, the JSON encoder does not know how to convert it into a JSON-compliant object. As a result, it throws the “Object of type ndarray is not JSON serializable” error. This error can occur when you try to serialize a Python object that contains an ndarray, such as a dictionary or a list.
Common Causes of the Error
The “Object of type ndarray is not JSON serializable” error can occur due to several reasons, including:
- Trying to serialize an object that contains an ndarray using the default JSON encoder
- Using the dump() or dumps() function from the json module to serialize an object that contains an ndarray
- Attempting to pass an ndarray as a parameter to a function that expects a JSON-compliant object
- Using a custom serializer that does not support ndarrays
How to Fix the Error
The “Object of type ndarray is not JSON serializable” error can be fixed by converting the ndarray into a JSON-compliant object. Here are some ways to do it:
Method 1: Using the tolist() Method
The easiest way to convert an ndarray into a JSON-compliant object is by using the tolist() method, which converts the ndarray into a list of nested lists. Here’s an example:
import numpy as npimport jsonndarray_obj = np.array([[1, 2, 3], [4, 5, 6]])list_obj = ndarray_obj.tolist()json_obj = json.dumps(list_obj)print(json_obj)In the above code, we first import the NumPy and JSON modules. We then create an ndarray object with two rows and three columns. We convert the ndarray into a list using the tolist() method and then use the dumps() function from the json module to serialize the list into a JSON string. Finally, we print the JSON string.
Method 2: Using a Custom Serializer
If you are working with a custom object that contains ndarrays, you can write your own serializer function that converts the ndarrays into JSON-compliant objects. Here’s an example:
import numpy as npimport jsonclass MyEncoder(json.JSONEncoder):def default(self, obj):if isinstance(obj, np.ndarray):return obj.tolist()return json.JSONEncoder.default(self, obj)ndarray_obj = np.array([[1, 2, 3], [4, 5, 6]])json_obj = json.dumps(ndarray_obj, cls=MyEncoder)print(json_obj)In the above code, we create a custom JSON encoder class called MyEncoder that inherits from the JSONEncoder class in the json module. We override the default() method of the JSONEncoder class to check if the object being serialized is an ndarray. If it is, we convert it into a list using the tolist() method. Otherwise, we call the default() method of the JSONEncoder class to handle the serialization of the object. We then create an ndarray object and use the dumps() function from the json module to serialize it into a JSON string using our custom encoder. Finally, we print the JSON string.
Method 3: Using the Object Hook
You can also use the object_hook parameter of the loads() function from the json module to convert ndarrays into JSON-compliant objects. Here’s an example:
import numpy as npimport jsondef convert_ndarray(obj):if '__ndarray__' in obj:return np.array(obj['__ndarray__'])return objndarray_obj = np.array([[1, 2, 3], [4, 5, 6]])json_obj = json.dumps({'data': ndarray_obj.tolist(), 'shape': ndarray_obj.shape, '__ndarray__': True})decoded_obj = json.loads(json_obj, object_hook=convert_ndarray)print(decoded_obj)In the above code, we define a function called convert_ndarray() that checks if the input object contains the string “__ndarray__”. If it does, we retrieve the ndarray data from the “__ndarray__” key and convert it into an ndarray object using the np.array() function. Otherwise, we return the input object. We then create an ndarray object and serialize it into a JSON string with an additional “shape” key and a “__ndarray__” key set to True. Finally, we use the loads() function from the json module to deserialize the JSON string with our custom object hook and print the decoded object.
Conclusion
The “Object of type ndarray is not JSON serializable” error can be frustrating when working with data in Python. However, it can be easily fixed by converting the ndarray into a JSON-compliant object using one of the methods we discussed in this article. By understanding the causes and solutions of this error, you can save time and avoid frustration when working with data in Python.