-1

我有一个用 python 为我编写的代码 json2csv 文件（我命名为 json2csv.py）。

import sys, json, csv

input = open(sys.argv[1])
json_array = json.load(input)
input.close()

item_data = json_array["items"]

if len(item_data) >= 1:
    first_item_id = item_data[0].keys()[0]
    columns = item_data[0][first_item_id].keys()

csv_file = open(sys.argv[2], "wb")
writer = csv.writer(csv_file)
# there is currently a known bug where column names are partially uppercase, this will be fixed soon. the "map(lambda x: x.lower(), columns)" fixes this issue in the mean time
writer.writerow(map(lambda x: x.lower(), columns)) 

# here .items() is a standard python function
for item_id, item_data in item_data[0].items():
    row = []
    for column_name in columns:
        if column_name.lower() == 'name_part': # lower required due to above issue
            row.append(" ".join(item_data[column_name]))
        else:
            row.append(item_data[column_name])
    writer.writerow(row) 

csv_file.close()

我的输入存储在文件 pd.json 中。pd.json 的内容是

{"status": "ok", "items": [{"work_phone_extension": null, "name_part": ["PATIENT", "TEST"], "residential_street_address_line_2": "", "referring_physician_first_name": "", "residential_street_address_line_1": "", "work_phone": "", "referring_physician_last_name": "", "residential_postal_or_zip_code": "", "referring_physician_code": "", "residence_phone": "416-", "health_card": "", "item_id": 1, "unique_vendor_id_sequence": 1, "cell_phone": null, "residential_city": "Toronto", "health_card_version": "", "residential_country_and_province_or_state": "CA_ON"}, {"work_phone_extension": null, "name_part": ["STEVE", "TEST"], "residential_street_address_line_2": "", "referring_physician_first_name": "Adam", "residential_street_address_line_1": "123 Fake St", "work_phone": "", "referring_physician_last_name": "Test", "residential_postal_or_zip_code": "M5E1Z8", "referring_physician_code": "123456", "residence_phone": "416-555-5555", "health_card": "", "item_id": 2, "unique_vendor_id_sequence": 2, "cell_phone": null, "residential_city": "Toronto", "health_card_version": "", "residential_country_and_province_or_state": "CA_ON"}]}

我正在执行命令

c:\python27\python.exe c:\Python27\Scripts\json2csv.py c:\Python27\Scripts\pd.json c:\Python27\Scripts\pd.txt

我不断收到错误

'Nonetype' object has no attribute keys.

有人可以在 first_item_id = item_data[0].keys()[0] 行指出 json2csv 中的错误吗

我厌倦了坐着处理这段代码。

此代码采用 pd.json 并将文件输出为 pd.txt

4

1 回答 1

1

你的错误应该是columns = item_data[0][first_item_id].keys()在线的，问题在于first_item_id，我认为你应该改变这个：

first_item_id = item_data[0].keys()[0]

对此：

first_item_id = item_data[0]['item_id']

但它仍然有一些问题，我认为代码不是为这个 json 结构编写的，我只是更改了另一行来修复它，我不确定这是正确的解决方案，但这是来自你当前代码的东西：

import sys, json, csv

input = open(sys.argv[1])
json_array = json.load(input)
input.close()

item_data = json_array["items"]
if len(item_data) >= 1:
    first_item_id = item_data[0]['item_id']
    columns = item_data[0].keys()

csv_file = open(sys.argv[2], "wb")
writer = csv.writer(csv_file)
# there is currently a known bug where column names are partially uppercase, this will be fixed soon. the "map(lambda x: x.lower(), columns)" fixes this issue in the mean time
writer.writerow(map(lambda x: x.lower(), columns)) 

# here .items() is a standard python function
for item in item_data:
    row = []
    for column_name in columns:
        if column_name.lower() == 'name_part': # lower required due to above issue
            row.append(" ".join(item[column_name]))
        else:
            row.append(item[column_name])
    writer.writerow(row) 

csv_file.close()
于 2013-03-29T23:48:44.677 回答