如何使用python或R将带有数组的JSON转换为CSV

我正在尝试将带有数组的JSON转换为CSV，但由于数组可以包含不同的内容，所以到目前为止我还没有找到解决方案

下面是一个JSON示例

[
{
    "name": "Doc 1",
    "description": "This is Document 1",
    "createdby": "User 1",
    "uid": "101",
    "created": "2020-01-01T12:00:00.000Z",
    "changed": "2020-01-01T13:00:00.000Z",
    "dim1": false,
    "dim2": false,
    "changedby": "User 1",
    "path": "/1/2/3"
},
{
    "name": "Doc 2",
    "description": "This is Document 2",
    "createdby": "User 1",
    "uid": "102",
    "created": "2020-01-01T12:00:00.000Z",
    "changed": "2020-01-01T13:00:00.000Z",
    "dim1": false,
    "dim2": false,
    "reference": [
        {
            "description": "Test1.csv",
            "uid": "9000.csv",
            "current": true
        }
    ],
    "changedby": "User 4",
    "path": "/1/2/4"
},
{
    "name": "Doc 3",
    "description": "This is Document 3",
    "createdby": "User 5",
    "uid": "105",
    "created": "2020-01-01T12:00:00.000Z",
    "changed": "2020-01-01T13:00:00.000Z",
    "dim1": false,
    "dim2": false,
    "reference": [
        {
            "description": "Test1.csv",
            "uid": "9000.csv",
            "current": true
        },
        {
            "description": "Test6.csv",
            "uid": "9005.csv",
            "current": true
        }
    ],
    "changedby": "User 4",
    "path": "/1/2/4"
},
{
    "name": "Doc 4",
    "description": "This is Document 4",
    "createdby": "User 2",
    "uid": "103",
    "created": "2020-01-01T12:00:00.000Z",
    "changed": "2020-01-01T13:00:00.000Z",
    "dim1": false,
    "dim2": false,
    "reference": [
        {
            "description": "Test2.sql",
            "uid": "9001.sql",
            "connection": {
                "type": "manual",
                "system": "SQL",
                "name": "Test2",
                "user": "sqlread1",
                "server": "server1.domain.com",
                "port": "1433",
                "sid": "300",
                "dim3": null
            },
            "current": false
        }
    ],
    "changedby": "User 4",
    "path": "/1/2/5"
},
{
    "name": "Doc 5",
    "description": "This is Document 5",
    "createdby": "User 3",
    "uid": "104",
    "created": "2020-01-01T12:00:00.000Z",
    "changed": "2020-01-01T13:00:00.000Z",
    "dim1": false,
    "dim2": false,
    "reference": [
        {
            "description": "Test3.sql",
            "uid": "9002.sql",
            "connection": {
                "type": "direct",
                "system": "SQL",
                "name": "Test3",
                "user": "sqlread2",
                "server": "server2.domain.com",
                "port": "1433",
                "sid": "301",
                "dim3": null
            },
            "current": false
        },
        {
            "description": "Test4.sql",
            "uid": "9003.sql",
            "connection": {
                "type": "manuel",
                "system": "SQL",
                "name": "Test4",
                "user": "sqlread3",
                "server": "server2.domain.com",
                "port": "1433",
                "sid": "302",
                "dim3": null
            },
            "current": false
        },
        {
            "description": "Test5.sql",
            "uid": "9004.sql",
            "connection": {
                "type": "direct",
                "system": "SQL",
                "name": "Test4",
                "user": "sqlread4",
                "server": "server2.domain.com",
                "port": "1433",
                "sid": "303",
                "dim3": null
            },
            "current": false
        },
        {
            "description": "Test6.csv",
            "uid": "9005.csv",
            "current": true
        }
    ],
    "changed": "User 4",
    "path": "/1/2/4"
}]

在这个JSON中，数组被称为“reference”，数组可能不存在，存在3个维度，或者11个维度，每个条目都有一个组“connection”，并混合了短条目和长条目

当我在python中使用这段代码时，我可以将JSON展平，但数组将位于一列中

import pandas as pd
from pandas.io.json import json_normalize
df=pd.read_json ('Sample.json')
print(df)

但我想要的应该是这样的：

对于数组中的每个附加条目，应复制该行，使整个数组内容位于相应的列中，但位于单独的行中

是否有可能使脚本具有通用性

谢谢！ 迈克

你也可以这样做：

---

是的，一切皆有可能！：）

你想做的事情通常被称为扁平化

---

你可以在这里找到一个例子：

他们将数组拆分为多列，我希望有多行，而不是多行。结果与我的json\u规范化代码相同-数组没有“分解”

import pandas as pd
import json

with open("yourjson.json", "r") as read_file:
    a = json.load(read_file)
df = pd.DataFrame(a)
print(df)
csv_data = df.to_csv('mynewjsonfile.csv', index = False)