如何将csv文件转换为python字典？

Question

我有一个巨大的 csv 文件，其中包含美国每个县的 COVID-19 病例和死亡信息。

为了让您大致了解此文件中包含的信息，这里是它的前 10 行：

date,county,state,fips,cases,deaths,confirmed_cases,confirmed_deaths,probable_cases,probable_deaths
2021-03-18,Autauga,Alabama,01001,6483,96,5557,85,926,11
2021-03-18,Baldwin,Alabama,01003,20263,295,14329,220,5934,75
2021-03-18,Barbour,Alabama,01005,2199,54,1225,37,974,17
2021-03-18,Bibb,Alabama,01007,2512,58,2031,35,481,23
2021-03-18,Blount,Alabama,01009,6371,129,4901,109,1470,20
2021-03-18,Bullock,Alabama,01011,1193,39,1059,29,134,10
2021-03-18,Butler,Alabama,01013,2069,66,1888,60,181,6
2021-03-18,Calhoun,Alabama,01015,14137,301,10608,242,3529,59
2021-03-18,Chambers,Alabama,01017,3460,113,1720,73,1740,40

无论如何，我想为此数据创建一个 Python 字典，其中每个键都是由州和县名称组成的元组，每个值都是整数列表：第一个 int 代表数字确诊病例数，第二个代表确诊死亡人数。

基本上，我想要这样的输出：

dic = {("state","county"):[confirmed_cases, confirmed_deaths]}

请确保排除标题。

如何为 csv 文件中的所有县生成一个类似于上述的 python 字典？请使用csv.reader。

此外，我需要找到某个特定州的所有已确认死亡人数的总和。例如，我将如何总结 'state' 为 'Alabama' 的所有行的 'confirmed death' 中的值？

编辑：我想出了解决问题第一部分的方法：

mydict = {}

with open(file_path, mode='r') as inp:
    reader = csv.reader(inp)
    next(reader,None)
    mydict = {tuple(row[1:3]):list(row[6:8]) for row in reader}

return mydict

你能帮我弄清楚如何根据这本词典总结某个州的已确认死亡人数吗？

Answer 1

我会建议一本字典。然后更容易按州或县获得总数。我使用了 2 个字典：dd 是 dict 的字典，deaths 是从 dd 字典构建的字典，用于按州获取死亡人数。

import csv
from collections import defaultdict

dd = defaultdict(dict)

with open('csv_11_09_01.csv', 'r') as f:
    reader = csv.DictReader(f)
    for record in reader:
        dd[ record['state'] ][ record['county'] ] = \
           [
               int(record['confirmed_cases']),
               int(record['confirmed_deaths'])
            ]

Answer 2

试试：

import csv

mydict = dict()
with open("test.csv") as inp:
    reader = csv.reader(inp)
    next(reader, None) #skip header
    mydict = {tuple(row[2:0:-1]): list(map(int, row[6:8])) for row in reader}

#total of all confirmed deaths in Alabama
>>> sum(v[1] for k, v in mydict.items() if k[0]=="Alabama")
890

test.csv：

date,county,state,fips,cases,deaths,confirmed_cases,confirmed_deaths,probable_cases,probable_deaths
2021-03-18,Autauga,Alabama,01001,6483,96,5557,85,926,11
2021-03-18,Baldwin,Alabama,01003,20263,295,14329,220,5934,75
2021-03-18,Barbour,Alabama,01005,2199,54,1225,37,974,17
2021-03-18,Bibb,Alabama,01007,2512,58,2031,35,481,23
2021-03-18,Blount,Alabama,01009,6371,129,4901,109,1470,20
2021-03-18,Bullock,Alabama,01011,1193,39,1059,29,134,10
2021-03-18,Butler,Alabama,01013,2069,66,1888,60,181,6
2021-03-18,Calhoun,Alabama,01015,14137,301,10608,242,3529,59
2021-03-18,Chambers,Alabama,01017,3460,113,1720,73,1740,40

Answer 3

我实际上会以不同的方式来做，虽然有点冗长，但对于查看代码的人来说更易读。

import csv
from collections import namedtuple

County = namedtuple("County", ["name", "cases", "deaths"])
reader = csv.DictReader(data)

for row in reader:
    state = row["state"]
    county = row["county"]
    record = County(county, int(row["confirmed_cases"]), int(row["confirmed_deaths"]))
    if state in states:
        states[state].append(record)
    else:
        states[state] = [record]

{'Alabama': [County(name='Autauga', cases=5557, deaths=85),
  County(name='Baldwin', cases=14329, deaths=220),
  County(name='Barbour', cases=1225, deaths=37),
  County(name='Bibb', cases=2031, deaths=35),
  County(name='Blount', cases=4901, deaths=109),
  County(name='Bullock', cases=1059, deaths=29),
  County(name='Butler', cases=1888, deaths=60),
  County(name='Calhoun', cases=10608, deaths=242),
  County(name='Chambers', cases=1720, deaths=73)]}

sum(county.deaths for county in states["Alabama"])
>> 890

如果您保持密钥简单（在本例中只是状态），则管理代码会更容易。如果您的数据更大，这也会更快，因为我们不必遍历字典中的元组键来获取我们想要的状态。

Answer 4

我认为 pandas 是最合适的解决方案：

import pandas as pd
df = pd.read_csv(file_path)
dict = df.set_index(['county','state'])[['confirmed_cases', 'confirmed_deaths']].apply(tuple, axis = 1).to_dict()
print(dict)

编辑

对于总和部分：

sum = df.groupby(['state'], as_index=False)['confirmed_cases', 'confirmed_deaths'].sum()
print(sum)