在python中將數據框轉換為嵌套的json

Question

我正在嘗試使用以下代碼將 df 轉換為嵌套的 json：

nested_json = (df.groupby(['prediction_probability','id','ts','prediction_value'], as_index=False)
             .apply(lambda x:x[[
                "first_create_date",  
                "create_date",
                "update_timestamp",
                 "revenue",
                 "col",
                 "x"]].to_dict('r'))
             .reset_index()
             .rename(columns={0:'features'})
             .to_json(orient='records'))

我的問題是嵌套的 dict (key ='features') 用方括號包裹。 如何避免方括號？ 我知道我可以將我的輸出視為字符串並替換方括號，但當然，這是一個不好的做法

輸出：

[
    {
        "pred": 0.50726,
        "id": "0030X00002qMwFrQAKxxxx",
        "ts": "2020-02-19T20:32:15.016586",
        "value": "A",
        "features": [
            {
                "first_create_date": 1582089665000,
                "create_date": 1582089665000,
                "update_timestamp": 1582142462000,
                "revenue": null,
                "col":"aaaa",
                "x": null
            }
        ]
    },
    {
        "pred": 0.50895,
        "id": "0030X00002qMvfHQASxxxxx",
        "ts": "2020-02-19T20:32:15.016586",
        "value": "A",
        "features": [
            {
                "first_create_date": 1582077985000,
                "create_date": 1582077985000,
                "update_timestamp": 1582142462000,
                "revenue": null,
                "col":"aaaa",
                "x": null
            }
        ]
    }
]

期望的輸出：

[
    {
        "pred": 0.50726,
        "id": "0030X00002qMwFrQAKxxxx",
        "ts": "2020-02-19T20:32:15.016586",
        "value": "A",
        "features": 
            {
                "first_create_date": 1582089665000,
                "create_date": 1582089665000,
                "update_timestamp": 1582142462000,
                "revenue": null,
                "col":"aaaa",
                "x": null
            }

    },
    {
        "pred": 0.50895,
        "id": "0030X00002qMvfHQASxxxxx",
        "ts": "2020-02-19T20:32:15.016586",
        "value": "A",
        "features": 
            {
                "first_create_date": 1582077985000,
                "create_date": 1582077985000,
                "update_timestamp": 1582142462000,
                "revenue": null,
                "col":"aaaa",
                "x": null
            }

    }
]

Answer 1

簡單的 dict 理解可以解決問題：假設您可以訪問一個形狀類似於輸出的嵌套 json 並將其稱為output 。 然后，要達到您想要的輸出，您唯一需要做的就是獲取features列表的第一個元素：

desired_output = [{k: v if k!='features' else v[0]} for x in output for k,v in x.items()]