如何从json文件中读取数据，并使用熊猫将其转换为csv？

Question

I am working on a data mining project. I need to read data from a json-format dataset which belongs to amazon.
The format of dataset is like this:
First I want to extract these rows:
[productName], [rating]
And after that I want to write the rows to a csv file with two columns named as productName and Rating. Is there any way to implement this by using pandas library?

Answer 1

With a subset of data, i have converted it into DF.Note that the data you have is not a json formatted data.

import pandas as pd
import json 
from collections import defaultdict
import re

f=open('inv.json')
text= f.readlines()
RowID=[]
result={}

for item in text:
    if item.startswith("###"):
        RowID=re.findall('\d+', item)
        result[RowID[0]]={}
    elif ":" in item:
        key,value =item.split(":",1)
        result[RowID[0]][key.strip()]=value.strip()
df= pd.DataFrame(result)
print df.transpose()

sample input

    #####1
[ID]:0
[ProductId]:0
[rating]:2.0

#####2
[ID]:1
[ProductId]:2
[rating]:3.0
[fullText]:It is a good
[weburl]:http://example.org:xx

output

       [ID] [ProductId]    [fullText] [rating]           [weburl]
1    0           0           NaN      2.0                NaN
2    1           2  It is a good      3.0  http://example.org:xx