如何用熊猫读取一组每个包含一个JSON对象的数组?

时间:2019-04-02 11:49:32

标签: json pandas

我正在尝试从文本文件读取到熊猫数据框。文本文件似乎是JSON的2D数组,我该如何读取?

[[{{'metric_name':'CPU','category':'A','data':'9','time_stamp':'2019-03-28 13:15:31'}],[ {'metric_name':'Disk','category':'B','data':'56','time_stamp':'2019-03-28 13:15:31'}]]

我希望将参数“ metric_name”,“ category”,“ data”,“ time_stamp”作为标题

1 个答案:

答案 0 :(得分:0)

这是一个解决方案:

import json
import pandas as pd

# load the file
raw_data = json.load(open('myfile.json'))

# raw_data contains a nested list, so convert it to a simple list :
data = [x[0] for x in raw_data]

# then create the dataframe
df = pd.DataFrame.from_records(data)

这里是data的内容。嵌套列表已转换为简单列表(假设每个数组有一条记录):

[{"category": "VM1",
      "data": "9",
      "metric_name": "CPU",
      "time_stamp": "2019-03-28 13:15:31"},
     {"category": "VM1",
      "data": "9",
      "metric_name": "CPU",
      "time_stamp": "2019-03-28 13:15:31"}]