Inverse of pandas json_normalize or json_denormalize – python pandas

As we all know pandas “json_normalize” which works great in taking a JSON Data, however, nested it is and convert’s it to the usable pandas dataframe.

pandas json_normalize documentation

Now If you want the reverse operation which takes that same Dataframe and convert back to originals JSON format, for example: for pushing data to elastic search DB or to store in Mongo DB or JSON File for Processing it later.

Here is a nice handy function which does exactly that, Thanks to Parsa T on StackOverflow for sharing this awesome utility function.

def make_formatted_dict(my_dict, key_arr, val):
    """
    Set val at path in my_dict defined by the string (or serializable object) array key_arr
    """
    current = my_dict
    for i in range(len(key_arr)):
        key = key_arr[i]
        if key not in current:
            if i == len(key_arr)-1:
                current[key] = val
            else:
                current[key] = {}
        else:
            if type(current[key]) is not dict:
                print("Given dictionary is not compatible with key structure requested")
                raise ValueError("Dictionary key already occupied")

        current = current[key]

    return my_dict


def df_to_formatted_json(df, sep="."):
    result = []
    for _, row in df.iterrows():
        parsed_row = {}
        for idx, val in row.iteritems():
            keys = idx.split(sep)
            parsed_row = make_formatted_dict(parsed_row, keys, val)

        result.append(parsed_row)
    return result

after backward converting you might need to drop empty columns from your generated JSON data, therefore, use below function that will remove empty columns:

import numpy as np

def remove_empties_from_dict(a_dict):
    new_dict = {}
    for k, v in a_dict.items():
        if isinstance(v, dict):
            v = remove_empties_from_dict(v)
        if v is not None and v is not np.nan and v is not "":
            new_dict[k] = v
    return new_dict


def make_formatted_dict(my_dict, key_arr, val):
    """
    Set val at path in my_dict defined by the string (or serializable object) array key_arr
    """
    current = my_dict
    for i in range(len(key_arr)):
        key = key_arr[i]
        if key not in current:
            if i == len(key_arr)-1:
                current[key] = val
            else:
                current[key] = {}
        else:
            if type(current[key]) is not dict:
                print("Given dictionary is not compatible with key structure requested")
                raise ValueError("Dictionary key already occupied")

        current = current[key]

    return remove_empties_from_dict(my_dict)


def df_to_formatted_json(df, sep="."):
    result = []
    for _, row in df.iterrows():
        parsed_row = {}
        for idx, val in row.iteritems():
            keys = idx.split(sep)
            parsed_row = make_formatted_dict(parsed_row, keys, val)

        result.append(parsed_row)
    return result

This Is Very Handy Utility Function.