将 json 嵌套到 csv - 通用方法

Question

我对 Python 非常陌生，我正在努力将嵌套的 json 文件转换为 cvs。为此，我首先加载json，然后将其转换为使用json_normalize 打印出漂亮输出的方式，然后使用pandas 包将标准化部分输出到cvs。

我的示例 json：

[{
 "_id": {
   "id": "123"
 },
 "device": {
   "browser": "Safari",
   "category": "d",
   "os": "Mac"
 },
 "exID": {
   "$oid": "123"
 },
 "extreme": false,
 "geo": {
   "city": "London",
   "country": "United Kingdom",
   "countryCode": "UK",
   "ip": "00.000.000.0"
 },
 "viewed": {
   "$date": "2011-02-12"
 },
 "attributes": [{
   "name": "gender",
   "numeric": 0,
   "value": 0
 }, {
   "name": "email",
   "value": false
 }],
 "change": [{
   "id": {
     "$id": "1231"
   },
   "seen": [{
     "$date": "2011-02-12"
   }]
 }]
}, {
 "_id": {
   "id": "456"
 },
 "device": {
   "browser": "Chrome 47",
   "category": "d",
   "os": "Windows"
 },
 "exID": {
   "$oid": "345"
 },
 "extreme": false,
 "geo": {
   "city": "Berlin",
   "country": "Germany",
   "countryCode": "DE",
   "ip": "00.000.000.0"
 },
 "viewed": {
   "$date": "2011-05-12"
 },
 "attributes": [{
   "name": "gender",
   "numeric": 1,
   "value": 1
 }, {
   "name": "email",
   "value": true
 }],
 "change": [{
   "id": {
     "$id": "1231"
   },
   "seen": [{
     "$date": "2011-02-12"
   }]
 }]
}]

使用以下代码（这里我排除了嵌套部分）：

import json
from pandas.io.json import json_normalize


def loading_file():
    #File path
    file_path = #file path here

    #Loading json file
    json_data = open(file_path)
    data = json.load(json_data)
    return data

#Storing avaliable keys
def data_keys(data):
    keys = {}
    for i in data:
        for k in i.keys():
            keys[k] = 1

    keys = keys.keys()

#Excluding nested arrays from keys - hard coded -> IMPROVE
    new_keys = [x for x in keys if
    x != 'attributes' and
    x != 'change']

    return new_keys

#Excluding nested arrays from json dictionary
def new_data(data, keys):
    new_data = []
    for i in range(0, len(data)):
        x = {k:v for (k,v) in data[i].items() if k in keys }
        new_data.append(x)
    return new_data

 def csv_out(data):
     data.to_csv('out.csv',encoding='utf-8')

def main():
     data_file = loading_file()
     keys = data_keys(data_file)
     table = new_data(data_file, keys)
     csv_out(json_normalize(table))

main()

我当前的输出如下所示：

| _id.id | device.browser | device.category | device.os |  ... | viewed.$date |
|--------|----------------|-----------------|-----------|------|--------------|
| 123    | Safari         | d               | Mac       | ...  | 2011-02-12   |
| 456    | Chrome 47      | d               | Windows   | ...  | 2011-05-12   |
|        |                |                 |           |      |              |

我的问题是我想将嵌套数组包含到 cvs 中，所以我必须将它们展平。我不知道如何使它通用，所以我在创建表时不使用字典 keys (numeric, id, name) 和 values。我必须使其具有普遍性，因为attributes 和change 中的键数。因此，我希望有这样的输出：

| _id.id | device.browser | ... | attributes_gender_numeric | attributes_gender_value | attributes_email_value | change_id | change_seen |
|--------|----------------|-----|---------------------------|-------------------------|------------------------|-----------|-------------|
| 123    | Safari         | ... | 0                         | 0                       | false                  | 1231      | 2011-02-12  |
| 456    | Chrome 47      | ... | 1                         | 1                       | true                   | 1231      | 2011-02-12  |
|        |                |     |                           |                         |                        |           |             |

提前感谢您！非常欢迎任何关于如何改进我的代码并使其更高效的提示。

Answer 1

感谢 Amir Ziai 的精彩博文，您可以找到 here 我设法以平面表格的形式输出我的数据。具有以下功能：

#Function that recursively extracts values out of the object into a flattened dictionary
def flatten_json(data):
    flat = [] #list of flat dictionaries
    def flatten(y):
        out = {}

        def flatten2(x, name=''):
            if type(x) is dict:
                for a in x:
                    if a == "name": 
                            flatten2(x["value"], name + x[a] + '_')
                    else:  
                        flatten2(x[a], name + a + '_')
            elif type(x) is list:
                for a in x:
                    flatten2(a, name + '_')
            else:
                out[name[:-1]] = x

        flatten2(y)
        return out

#Loop needed to flatten multiple objects
    for i in range(len(data)):
        flat.append(flatten(data[i]).copy())

    return json_normalize(flat) 

我知道由于名称-值 if 语句，它不是完全可推广的。但是，如果删除了创建名称-值字典的豁免，则该代码可以与其他嵌入式数组一起使用。

Answer 2

几周前，我有一项任务是将带有嵌套键和值的 json 转换为 csv 文件。对于此任务，必须正确处理嵌套键以连接要用作值的唯一标头。结果是下面的代码，也可以找到here。

def get_flat_json(json_data, header_string, header, row):
    """Parse json files with nested key-vales into flat lists using nested column labeling"""
    for root_key, root_value in json_data.items():
        if isinstance(root_value, dict):
            get_flat_json(root_value, header_string + '_' + str(root_key), header, row)
        elif isinstance(root_value, list):
            for value_index in range(len(root_value)):
                for nested_key, nested_value in root_value[value_index].items():
                    header[0].append((header_string +
                                      '_' + str(root_key) +
                                      '_' + str(nested_key) +
                                      '_' + str(value_index)).strip('_'))
                    if nested_value is None:
                        nested_value = ''
                    row[0].append(str(nested_value))
        else:
            if root_value is None:
                root_value = ''
            header[0].append((header_string + '_' + str(root_key)).strip('_'))
            row[0].append(root_value)
    return header, row

这是一种基于经济学家对此问题的回答的更通用的方法。

Answer 3

以下代码处理了一个杂乱无章的 json 文件 字典和列表相互之间 7 层深：

    import csv, json, os
    def parse_json(data):
        a_dict_accum = {}
        for key, val in data.items():
            print("key, val = ", key, val)
            output.writerow([key])
            output.writerow([val])
            if isinstance(val, dict):
                for a_key, a_val in val.items():
                    print("a_key, a_val = ", a_key, a_val)
                    output.writerow([a_key])
                    output.writerow([a_val])
                    a_dict_accum.update({a_key:a_val})
                print("a_dict_accum = ", a_dict_accum)
                parse_json(a_dict_accum)
            elif isinstance(val, list):
                print("val_list = ", val)
                for a_list in val:
                     print("a_list = ", a_list)
                     output.writerow([a_list])
                     if isinstance(a_list, dict):
                         for a_key, a_val in a_list.items():
                             print("a_key, a_val = ", a_key, a_val)
                             output.writerow([a_key])
                             output.writerow([a_val])    
                             a_dict_accum.update({a_key:a_val})
                         print("a_dict_accum = ", a_dict_accum)
                         parse_json(a_dict_accum)
    os.chdir('C://Users/Robert/viirs/20200217_output')
    fileInput = 'night_lights_points.json'
    fileOutput = 'night_lights_points.csv'
    inputFile = open(fileInput) #open json file
    outputFile = open(fileOutput, 'w', newline='') #load csv file
    data = json.load(inputFile) #load json content
    output = csv.writer(outputFile) #create a csv.writer
    output = parse_json(data)
    inputFile.close() #close the input file
    outputFile.close() #close the output file