JSON读入Pandas DataFrame部分数据丢失？

Question

I want to do some analysis on data I have collected from a game study. We store a Timestamp, the Input Type and then the Metadata for the respective rounds that were played. We store it as a JSON, and I wanted to load it into a python script to generate some nice graphics with matplotlib. To use Pandas, I wanted to convert it into a.CSV for a dataframe, however some data seems to go missing when printing the DF.

    with open('/Users/me/Downloads/databackup.json', 'r') as f:
        data = json.loads(f.read())
    
    multiple_level_data = pd.json_normalize(data, record_path=['gameList'],
                                            meta=[], meta_prefix='config_params_',
                                            record_prefix='dbscan_')
    
    multiple_level_data.to_csv('GameData.csv', index=False)
    df = pd.read_csv("GameData.csv")

This is what I use to convert the JSON into a CSV. Now, we create a new Timestamp every time the player reached a score of 750 in his last x rounds. When theres only one Round, the data for that Timestamp shows up, but when theres two or more Rounds per Timestamp, the respective Data of those rounds does not show up in my df. Did I choose the wrong record_path or am I using the wrong method to convert this?

{
    "gameList": [
        {
            "startingTime": "20230125204032",
            "inputType": "joyStick",
            "Rounds": [
                {
                    "durationSeconds": 128,
                    "score": 492,
                    "platformCount": {
                        "normalPlatforms": 60,
                        "movingPlatforms": 41,            #this loads in
                        "powPlatforms": 5,
                        "normalEnemies": 5,
                        "movingEnemies": 8
                    }
                },
                {
                    "durationSeconds": 62,
                    "score": 258,
                    "platformCount": {
                        "normalPlatforms": 35,
                        "movingPlatforms": 23,             #this doesn't
                        "powPlatforms": 3,
                        "normalEnemies": 2,
                        "movingEnemies": 5
                    }
                }
            ]
        },

Answer 1

If you want to follow the path down the levels you need to use a list. If you want multiple items in the same level as in the case of meta you need to map the levels in a list of lists.

df = pd.json_normalize(
    data=data,
    record_path=["gameList", "Rounds"],
    meta_prefix="config_params_",
    record_prefix="dbscan_",
    meta=[["gameList", "startingTime"], ["gameList", "inputType"]]
)

Output:

   dbscan_durationSeconds  dbscan_score  dbscan_platformCount.normalPlatforms  dbscan_platformCount.movingPlatforms  dbscan_platformCount.powPlatforms  dbscan_platformCount.normalEnemies  dbscan_platformCount.movingEnemies config_params_gameList.startingTime config_params_gameList.inputType
0                     128           492                                    60                                    41                                  5                                   5                                   8                      20230125204032                         joyStick
1                      62           258                                    35                                    23                                  3                                   2                                   5                      20230125204032                         joyStick