如何将多个 json 文件读入 pandas 数据框？

Question

I'm having a hard time loading multiple line delimited JSON files into a single pandas dataframe. This is the code I'm using:

import os, json
import pandas as pd
import numpy as np
import glob
pd.set_option('display.max_columns', None)

temp = pd.DataFrame()

path_to_json = '/Users/XXX/Desktop/Facebook Data/*' 

json_pattern = os.path.join(path_to_json,'*.json')
file_list = glob.glob(json_pattern)

for file in file_list:
    data = pd.read_json(file, lines=True)
    temp.append(data, ignore_index = True)

It looks like all the files are loading when I look through file_list , but cannot figure out how to get each file into a dataframe. There are about 50 files with a couple lines in each file.

Answer 1

Change the last line to:

temp = temp.append(data, ignore_index = True)

The reason we have to do this is because the append doesn't happen in place. The append method does not modify the data frame. It just returns a new data frame with the result of the append operation.

Edit:

Since writing this answer I have learned that you should never use DataFrame.append inside a loop because it leads to quadratic copying (see this answer ).

What you should do instead is first create a list of data frames and then use pd.concat to concatenate them all in a single operation. Like this:

dfs = [] # an empty list to store the data frames
for file in file_list:
    data = pd.read_json(file, lines=True) # read data frame from json file
    dfs.append(data) # append the data frame to the list

temp = pd.concat(dfs, ignore_index=True) # concatenate all the data frames in the list.

This alternative should be considerably faster.

Answer 2

If you need to flatten the JSON, Juan Estevez's approach won't work as is. Here is an alternative :

import pandas as pd

dfs = []
for file in file_list:
    with open(file) as f:
        json_data = pd.json_normalize(json.loads(f.read()))
    dfs.append(json_data)
df = pd.concat(dfs, sort=False) # or sort=True depending on your needs

Or if your JSON are line-delimited (not tested) :

import pandas as pd

dfs = []
for file in file_list:
    with open(file) as f:
        for line in f.readlines():
            json_data = pd.json_normalize(json.loads(line))
            dfs.append(json_data)
df = pd.concat(dfs, sort=False) # or sort=True depending on your needs

Answer 3

from pathlib import Path
import pandas as pd

paths = Path("/home/data").glob("*.json")
df = pd.DataFrame([pd.read_json(p, typ="series") for p in paths])```

Answer 4

Maybe you should state, if the json files are created themselves with pandas pd.to_json() or in another way. I used data which was not created with pd.to_json() and I think it is not pssible to use pd.read_json() in my case. Instead, I programmed a customized for-each loop approach to write everything to the DataFrames

Answer 5

I combined Juan Estevez's answer with glob. Thanks a lot.

import pandas as pd
import glob

def readFiles(path):
    files = glob.glob(path)
    dfs = [] # an empty list to store the data frames
    for file in files:
        data = pd.read_json(file, lines=True) # read data frame from json file
        dfs.append(data) # append the data frame to the list

    df = pd.concat(dfs, ignore_index=True) # concatenate all the data frames in the list.
    return df