處理巢狀資料欄位
使用 pandas 重塑資料
Maria Eugenia Inzaugarat
Data Scientist
複習
- 使用
json_normalize()將巢狀 JSON 讀入 DataFrame。
欄位中的巢狀資料
writers = ["Mary Shelley", "Ernest Hemingway"] books = ["{'title': 'Frankenstein', 'year': 1818}", "{'title': 'The Old Man and the Sea', 'year': 1951}"]collection = pd.DataFrame( )
欄位中的巢狀資料
writers = ["Mary Shelley", "Ernest Hemingway"] books = ["{'title': 'Frankenstein', 'year': 1818}", "{'title': 'The Old Man and the Sea', 'year': 1951}"]collection = pd.DataFrame(dict( ))
欄位中的巢狀資料
writers = ["Mary Shelley", "Ernest Hemingway"] books = ['{"title": "Frankenstein", "year": "1818"}', '{"title": "The Old Man and the Sea", "year":"1951"}']collection = pd.DataFrame(dict(writers=writers, books=books))collection
writers books
0 Mary Shelley {'title': 'Frankenstein', 'year': 1818}
1 Ernest Hemingway {'title': 'The Old Man and the Sea', 'year': 1951}
轉換巢狀資料
import jsonbooks = collection['books']
轉換巢狀資料
import jsonbooks = collection['books'].apply( )
轉換巢狀資料
import jsonbooks = collection['books'].apply(json.loads)
轉換巢狀資料
import jsonbooks = collection['books'].apply(json.loads).apply(pd.Series)books
title year
0 Frankenstein 1818
1 The Old Man and the Sea 1951
再串回去
collection = collection.drop(columns='books')pd.concat([collection, books], axis=1)
writers title year
0 Mary Shelley Frankenstein 1818
1 Ernest Hemingway The Old Man and the Sea 1951
序列化巢狀資料
import json books = collection['books'].apply(json.loads)
序列化巢狀資料
import json books = collection['books'].apply(json.loads).to_list()books_dump = json.dumps(books)new_books = pd.read_json(books_dump)new_books
title year
0 Frankenstein 1818
1 The Old Man and the Sea 1951
序列化巢狀資料
pd.concat([collection['writers'], new_books], axis=1)
writers title year
0 Mary Shelley Frankenstein 1818
1 Ernest Hemingway The Old Man and the Sea 1951
一起來練習吧！
使用 pandas 重塑資料
