1. ホーム
  2. python

[解決済み】1行以上のJSONを含むファイルをPandasに読み込む

2022-01-29 07:07:45

質問

Python pandas (0.14.0) のデータフレームに JSON ファイルを読み込もうとしています。以下はJSONファイルの1行目です。

{"votes": {"funny": 0, "useful": 0, "cool": 0}, "user_id": "P_Mk0ygOilLJo4_WEvabAA", "review_id": "OeT5kgUOe3vcN7H6ImVmZQ", "stars": 3, "date": "2005-08-26", "text": "This is a pretty typical cafe.  The sandwiches and wraps are good but a little overpriced and the food items are the same.  The chicken caesar salad wrap is my favorite here but everything else is pretty much par for the course.", "type": "review", "business_id": "Jp9svt7sRT4zwdbzQ8KQmw"}

私は以下のようなことをしようとしています。 df = pd.read_json(path) .

次のようなエラーが発生します(完全なトレースバックあり)。

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/d/anaconda/lib/python2.7/site-packages/pandas/io/json.py", line 198, in read_json
    date_unit).parse()
  File "/Users/d/anaconda/lib/python2.7/site-packages/pandas/io/json.py", line 266, in parse
    self._parse_no_numpy()
  File "/Users/d/anaconda/lib/python2.7/site-packages/pandas/io/json.py", line 483, in _parse_no_numpy
    loads(json, precise_float=self.precise_float), dtype=None)
ValueError: Trailing data

とは何ですか? Trailing data のエラーは?データフレームに読み込むにはどうしたらいいですか?

いくつかの提案に従って、以下は.jsonファイルの数行です。

{"votes": {"funny": 0, "useful": 0, "cool": 0}, "user_id": "P_Mk0ygOilLJo4_WEvabAA", "review_id": "OeT5kgUOe3vcN7H6ImVmZQ", "stars": 3, "date": "2005-08-26", "text": "This is a pretty typical cafe.  The sandwiches and wraps are good but a little overpriced and the food items are the same.  The chicken caesar salad wrap is my favorite here but everything else is pretty much par for the course.", "type": "review", "business_id": "Jp9svt7sRT4zwdbzQ8KQmw"}
{"votes": {"funny": 0, "useful": 0, "cool": 0}, "user_id": "TNJRTBrl0yjtpAACr1Bthg", "review_id": "qq3zF2dDUh3EjMDuKBqhEA", "stars": 3, "date": "2005-11-23", "text": "I agree with other reviewers - this is a pretty typical financial district cafe.  However, they have fantastic pies.  I ordered three pies for an office event (apple, pumpkin cheesecake, and pecan) - all were delicious, particularly the cheesecake.  The sucker weighed in about 4 pounds - no joke.\n\nNo surprises on the cafe side - great pies and cakes from the catering business.", "type": "review", "business_id": "Jp9svt7sRT4zwdbzQ8KQmw"}
{"votes": {"funny": 0, "useful": 0, "cool": 0}, "user_id": "H_mngeK3DmjlOu595zZMsA", "review_id": "i3eQTINJXe3WUmyIpvhE9w", "stars": 3, "date": "2005-11-23", "text": "Decent enough food, but very overpriced. Just a large soup is almost $5. Their specials are $6.50, and with an overpriced soda or juice, it's approaching $10. A bit much for a cafe lunch!", "type": "review", "business_id": "Jp9svt7sRT4zwdbzQ8KQmw"}

私が使用しているこの.jsonファイルには、仕様通り1行に1つのJSONオブジェクトが含まれています。

を試してみました。 jsonlint.com のサイトで提案されたとおり、以下のようなエラーが発生します。

Parse error on line 14:
...t7sRT4zwdbzQ8KQmw"}{    "votes": {
----------------------^
Expecting 'EOF', '}', ',', ']'

解決方法は?

Pandasのバージョン0.19.0からは lines のようなパラメータがあります。

import pandas as pd

data = pd.read_json('/path/to/file.json', lines=True)