根据日期计算给定文本文件中每个单词的出现次数

Question

我目前有一个文本文件，其中包含某人发送消息时的时间戳和姓名。请看下文：Attachment

8/29/19, 2:03 PM - Michael: ...
8/29/19, 3:05 PM - Frank: ...
8/29/19, 4:01 PM - Tom: ...
8/29/19, 5:26 PM - Amy: ...
8/29/19, 6:46 PM - Tom: ...
8/29/19, 7:24 PM - Frank: ...
8/29/19, 9:55 PM - Amy: ...
8/30/19, 11:35 AM - Frank: ...
8/30/19, 12:39 PM - Johnny: ...
9/3/19, 1:18 AM - Frank: ...
9/3/19, 2:23 AM - Frank: ...
9/3/19, 3:16 PM - Frank: ...
9/3/19, 4:53 PM - Johnny: ...
9/4/19, 9:01 AM - Frank: ...
9/4/19, 11:45 AM - Frank: ...
9/4/19, 1:04 PM - Johnny: ...
9/4/19, 1:42 PM - Johnny: ...
9/4/19, 2:03 PM - Amy: ...
9/4/19, 4:12 PM - Johnny: ...
9/4/19, 6:27 PM - Amy: ...
9/4/19, 9:08 PM - Johnny: ...
.   .      .
.   .      .
.   .      .

我想根据python中的日期计算一个人发送消息的发生次数。我想输出以下内容：Attachment

                 Michael  Frank   Tom    Amy     Johnny

      8/29/2019     1        2      2     2       0
      8/30/2019     0        1      0     0       1
      8/31/2019     0        0      0     0       0
      9/1/2019      0        0      0     0       0
      9/2/2019      0        0      0     0       0
      9/3/2019      0        3      0     0       1
      9/4/2019      0        2      0     2       4
      9/5/2019                  
      9/6/2019                  
      9/7/2019                  
      9/8/2019                  

第一次发帖，如果格式不正确，请见谅。非常感谢。

Answer 1

您可以使用 pandas 来提供帮助：

from io import StringIO
import pandas as pd

txtfile=StringIO("""8/29/19, 2:03 PM - Michael: ...
8/29/19, 3:05 PM - Frank: ...
8/29/19, 4:01 PM - Tom: ...
8/29/19, 5:26 PM - Amy: ...
8/29/19, 6:46 PM - Tom: ...
8/29/19, 7:24 PM - Frank: ...
8/29/19, 9:55 PM - Amy: ...
8/30/19, 11:35 AM - Frank: ...
8/30/19, 12:39 PM - Johnny: ...
9/3/19, 1:18 AM - Frank: ...
9/3/19, 2:23 AM - Frank: ...
9/3/19, 3:16 PM - Frank: ...
9/3/19, 4:53 PM - Johnny: ...
9/4/19, 9:01 AM - Frank: ...
9/4/19, 11:45 AM - Frank: ...
9/4/19, 1:04 PM - Johnny: ...
9/4/19, 1:42 PM - Johnny: ...
9/4/19, 2:03 PM - Amy: ...
9/4/19, 4:12 PM - Johnny: ...
9/4/19, 6:27 PM - Amy: ...
9/4/19, 9:08 PM - Johnny: ...""")

df = pd.read_csv(txtfile, sep=',|-|:', header=None, index_col=[0], engine='python')
df_out = df[3].str.get_dummies().sum(level=0)
print(df_out)

输出：

          Amy   Frank   Johnny   Michael   Tom
0                                             
8/29/19     2       2        0         1     2
8/30/19     0       1        1         0     0
9/3/19      0       3        1         0     0
9/4/19      2       2        4         0     0

Answer 2

通过只在文件中迭代一次来做到这一点的一种方法是在 defaultdict 的帮助下：

from collections import defaultdict

occurrences = defaultdict(lambda: defaultdict(int))

with open('filename.txt', 'r') as f:
    for line in f.readlines():
        date = line.split(', ')[0]
        name = line.split(' - ')[1].split(': ')[0]
        occurrences[date][name] += 1

出现次数将有以下数据：

8/29/19: {'Michael': 1, 'Frank': 2, 'Tom': 2, 'Amy': 2}
8/30/19: {'Frank': 1, 'Johnny': 1}
9/3/19: {'Frank': 3, 'Johnny': 1}
9/4/19: {'Frank': 2, 'Johnny': 4, 'Amy': 2}

编辑：这将打印出 OP 想要的确切输出：

from collections import defaultdict
from datetime import datetime, timedelta

occurrences = defaultdict(lambda: defaultdict(int))

with open('filename.txt', 'r') as f:
    lines = f.readlines()
    start_date = lines[0].split(' - ')[0]
    start_date = datetime.strptime(start_date, '%m/%d/%y, %I:%M %p')
    end_date = lines[-1].split(' - ')[0]
    end_date = datetime.strptime(end_date, '%m/%d/%y, %I:%M %p')

    dates = []

    for n in (range(int((end_date - start_date).days))):
        single_date = start_date + timedelta(n)
        dates.append(single_date.date())

    authors = set()

    for line in lines:
        name = line.split(' - ')[1].split(': ')[0]
        authors.add(name)
        date = line.split(' - ')[0]
        date = datetime.strptime(date, '%m/%d/%y, %I:%M %p').date()
        occurrences[date][name] += 1

    print('\t\t', end='')
    for name in authors:
        print (name, end='\t')
    print()

    for date in dates:
        print(date.strftime('%m/%d/%y'), end='\t')
        for name in authors:
            print(occurrences[date][name], end='\t')
        print()

这个解决方案有改进的余地，因为它完全无视性能。