在字典数组中分组 csv 数据 - Python答案

【问题标题】：Group csv data in a array of dictionaries - Python在字典数组中分组 csv 数据 - Python
【发布时间】：2021-07-29 15:27:23
【问题描述】：

我有一个这样的 CSV 文件：（userId、movieId、score）并按 userId 排序

user1,movie1,0.1
user1,movie2,0.2
user2,movie2,0.4
user2,movie1,0.2

我想将它们分组到这样的字典数组中：

[
   {
      "userId":"user1",
      "scores":[
         {
            "movieId":"movie1",
            "score":0.1
         },
         {
            "movieId":"movie2",
            "score":0.2
         },
         
      ]
   },
   {
      "userId":"user2",
      "scores":[
         {
            "movieId":"movie2",
            "score":0.4
         },
         {
            "movieId":"movie1",
            "score":0.2
         }
      ]
   }
]

这是我使用 python 的尝试，但它不起作用

def get_body(batch):
    
    result = []
    record = {}
    scores = []
   
    for row in batch:
        if 'userId' in record and record['userId'] != row[0]:
            result.append({'userId': record['userId'], 'scores': scores})
            record = {}
            scores = []
        
        if 'userId' not in record:
            record['userId'] = row[0]

        scores.append({'movieId': row[1], 'score': float(row[2])})
        
    return result

另外，我没有使用 pandas 作为替代品，非常感谢您的帮助

【问题讨论】：

标签： python arrays dictionary

【解决方案1】：

仅使用内置 csv 模块：

import csv
import json

out = {}
with open("your_file.csv", "r") as f_in:
    reader = csv.reader(f_in)
    for row in reader:
        out.setdefault(row[0], []).append(
            {"movieId": row[1], "score": float(row[2])}
        )

out = [{"userId": k, "scores": v} for k, v in out.items()]
# pretty print:
print(json.dumps(out, indent=4))

打印：

[
    {
        "userId": "user1",
        "scores": [
            {
                "movieId": "movie1",
                "score": 0.1
            },
            {
                "movieId": "movie2",
                "score": 0.2
            }
        ]
    },
    {
        "userId": "user2",
        "scores": [
            {
                "movieId": "movie2",
                "score": 0.4
            },
            {
                "movieId": "movie1",
                "score": 0.2
            }
        ]
    }
]

【讨论】：