CSV 到多值字典？

Question

                Title_1                                Title_2         Type
 He heard it from space  A quick story about sounds from space      Fiction
    The end of all time       A sad poem about the end of time  Non-Fiction
  The perfect beginning               A story about friendship  Non-Fiction

我正在尝试计算所有小说、非小说类型并计算 Title_1 和 Title_2 中相应类型的单词数。

我想要的输出是：

Type         Count  Num-Words  
Non-Fiction   2       20
Fiction       1       12

这是我目前所拥有的：

fopen =  open(file_name, 'r')
fhand = csv.reader(fopen)
next(fhand)
category_sum = dict()
for row in fhand:
    col_0=len(row[0].split())
    col_1=len(row[1].split())
    print( col_1 + col_1)
    if row[2] in category_sum.keys():
        category_sum[row[2]]+=1
    else:
        category_sum[row[2]]=1

我可以在一本不错的字典中获得类型的总数，但我似乎无法弄清楚如何将字数分配给适当的类型作为字典中的值。

有什么想法吗？

Answer 1

您可以将字典保存为其中一个键为Count 和另一个键为Num-Words 的值。所以你的字典值赋值可能看起来像：

# num_of_words = 
if row[2] in category_sum.keys():
    category_sum[row[2]]['Count']+=1
    category_sum[row[2]]['Num-Words']+=num_of_words
else:
    category_sum[row[2]]={}

Answer 2

你可以这样做：

import csv

file_name = 'book_titles.csv'

with open(file_name, 'r', newline='') as fopen:
    reader = csv.reader(fopen)
    next(reader)  # Skip header.
    category_sum = {}
    for row in reader:
        category_sum[row[2]] = category_sum.get(row[2], 0) + 1

print(category_sum)  # -> {'Fiction': 1, 'Non-Fiction': 2}

Answer 3

使用pandas:

• 创建数据框
• 合并两个标题，按空格分开并计算split创建的列表中的单词
• groupby on Type，然后聚合 count 和 sum 函数。
  • reset_index 和 rename 以获得所需的确切形式。

import pandas as pd

# read the file in
df = pd.read_csv('file.csv')

                Title_1                                Title_2         Type
 He heard it from space  A quick story about sounds from space      Fiction
    The end of all time       A sad poem about the end of time  Non-Fiction
  The perfect beginning               A story about friendship  Non-Fiction

# count the words in Title_1 & Title_2
df['num_words'] = df[['Title_1', 'Title_2']].apply(lambda x: len(f'{x[0]} {x[1]}'.split()), axis=1)

                Title_1                                Title_2         Type  num_words
 He heard it from space  A quick story about sounds from space      Fiction         12
    The end of all time       A sad poem about the end of time  Non-Fiction         13
  The perfect beginning               A story about friendship  Non-Fiction          7

# create your desired output
test = df[['Type', 'num_words']].groupby('Type')['num_words'].agg(['count', 'sum']).reset_index().rename(columns={'count': 'Count', 'sum': 'Num-words'})

        Type  Count  Num-words
     Fiction      1         12
 Non-Fiction      2         20

• 只需 3 行代码即可获得所需的输出
• 使用数据框中的数据，如果需要，您可以更轻松地执行其他类型的文本分析（例如 Text analysis: finding the most common word in a column using python）

在dict 中获取输出：

test.to_dict('list')

>>> {'Type': ['Fiction', 'Non-Fiction'], 'Count': [1, 2], 'Num-words': [12, 20]}

Answer 4

这是我最终使用的：

fhand = csv.reader(fopen)
next(fhand)
category_sum = dict()
word_sum = dict()

for row in fhand:
    num_words = len(row[0].split(" ")) + len(row[1].split(" "))
    if row[2] in category_sum.keys():
        category_sum[row[2]]+=1
        word_sum[row[2]]+=num_words
    else:
        category_sum[row[2]]=1
        word_sum[row[2]]=num_words



combined = {key:[category_sum[key],word_sum[key]] for key in category_sum}   
#print(combined)
print("Category | # Titles | # of Words\n---------------------------------")
for key in combined:
    print("{}   |   {}  |   {}  ".format(key,combined[key][0],combined[key][1]))