pandas-从json数据创建数据框架,具体包括哪些列

时间:2018-08-17 12:15:19

标签: python pandas

我当前正在使用以下命令创建数据框

getting data from url
...
devices = get_device_data.json()
device_data = devices["data"]
p_dev = pd.DataFrame(device_data)

但是json字典中有60列,我只希望其中两列,有没有一种方法可以指定在创建数据框时要包括哪些列,或者我可以以任何方式达到预期的结果?

谢谢

编辑:一些示例数据删除了某些列。我实际上只想要id和hostname列

[
    {
        "id": 474378238
        "account": "https: //www.****.com/api/v2/accounts/38021/",
        "bsid": None,
        "carrier": "BOB",
        "carrier_id": "BOB BOB",
        "channel": None,
        "connection_state": "connected",
        "gsn": "356853050758871",
        "homecarrid": "BOB",
        "hostname": "BOB1345345",
        "is_asset": True,
        "is_gps_supported": True,
        "is_upgrade_available": False,
        "is_upgrade_supported": True,
        "ltebandwidth": "20 MHz",
        "mac": None,
        "serial": "356853050758871",
        "service_type": "LTE",
        "ssid": None,
        "summary": "connected",
        "txchannel": "19667",
        "type": "mdm",
        "uid": "f5a8da8f",
        "updated_at": "2018-08-17T11:19:57.019938+00:00",
        "uptime": 86412.8558200002,
    },
    {
        "id": 5674657356
        "account": "https: //www.****.com/api/v2/accounts/38021/",
        "bsid": None,
        "carrier": "BOB",
        "carrier_id": "BOB BOB",
        "channel": None,
        "connection_state": "connected",
        "gsn": "356853050758871",
        "homecarrid": "BOB",
        "hostname": "BOB10765",
        "is_asset": True,
        "is_gps_supported": True,
        "is_upgrade_available": False,
        "is_upgrade_supported": True,
        "ltebandwidth": "20 MHz",
        "mac": None,
        "serial": "356853050758871",
        "service_type": "LTE",
        "ssid": None,
        "summary": "connected",
        "txchannel": "19667",
        "type": "mdm",
        "uid": "f5a8da8f",
        "updated_at": "2018-08-17T11:19:57.019938+00:00",
        "uptime": 86412.8558200002,
    },
    {
        "id": 5674657465
        "account": "https: //www.****.com/api/v2/accounts/38021/",
        "bsid": None,
        "carrier": "BOB",
        "carrier_id": "BOB BOB",
        "channel": None,
        "connection_state": "connected",
        "gsn": "356853050758871",
        "homecarrid": "BOB",
        "hostname": "BOB10453453",
        "is_asset": True,
        "is_gps_supported": True,
        "is_upgrade_available": False,
        "is_upgrade_supported": True,
        "ltebandwidth": "20 MHz",
        "mac": None,
        "serial": "356853050758871",
        "service_type": "LTE",
        "ssid": None,
        "summary": "connected",
        "txchannel": "19667",
        "type": "mdm",
        "uid": "f5a8da8f",
        "updated_at": "2018-08-17T11:19:57.019938+00:00",
        "uptime": 86412.8558200002,
    },
    {
        "id": 9756756756
        "account": "https: //www.****.com/api/v2/accounts/38021/",
        "bsid": None,
        "carrier": "BOB",
        "carrier_id": "BOB BOB",
        "channel": None,
        "connection_state": "connected",
        "gsn": "356853050758871",
        "homecarrid": "BOB",
        "hostname": "BOB100133",
        "is_asset": True,
        "is_gps_supported": True,
        "is_upgrade_available": False,
        "is_upgrade_supported": True,
        "ltebandwidth": "20 MHz",
        "mac": None,
        "serial": "356853050758871",
        "service_type": "LTE",
        "ssid": None,
        "summary": "connected",
        "txchannel": "19667",
        "type": "mdm",
        "uid": "f5a8da8f",
        "updated_at": "2018-08-17T11:19:57.019938+00:00",
        "uptime": 86412.8558200002,
    },  
]       

2 个答案:

答案 0 :(得分:1)

使用list comprehensiondict comprehension来按列名称进行过滤:

L是输入数据列表

device_data = [{k: v for k, v in x.items() if k in ['type','id']} for x in L]
print (device_data)
[{'id': 474378238, 'type': 'mdm'}, {'id': 5674657356, 'type': 'mdm'}, 
 {'id': 5674657465, 'type': 'mdm'}, {'id': 9756756756, 'type': 'mdm'}]

df = pd.DataFrame(device_data)
print (df)
           id type
0   474378238  mdm
1  5674657356  mdm
2  5674657465  mdm
3  9756756756  mdm

答案 1 :(得分:0)

您可以通过使用列标题从数据框中删除那些列来实现。

示例:p_dev.drop(['Column_1','Column_2','column_2'], axis = 1, inplace = True)

另一种方法是只将列表中需要的列标题写入列表,然后覆盖现有数据框

示例:

col_list = ['Column_1','Column_3']

p_def = p_def [col_list]