Question

我目前有一个字典，该字典存储用户的ID作为键，并将用户执行的事件存储为元组列表。每个元组都包含事件执行的日期和事件本身。

这是字典的摘录：

{
 '56d306892fcf7d8a0563b488bbe72b0df1c42f8b62edf18f68a180eab2ca7dc5': 
  [('2018-10-24T08:30:12.761Z', 'booking_initialized')],
 'ac3406118670ef98ee2e3e76ab0f21edccba7b41fa6e4960eea10d2a4d234845': 
  [('2018-10-20T14:12:35.088Z', 'visited_hotel'), ('2018-10-20T14:17:38.521Z', 
 'visited_hotel'), ('2018-10-20T14:16:41.968Z', 'visited_hotel'), ('2018-10- 
 20T13:39:36.064Z', 'search_hotel'), ('2018-10-20T13:47:03.086Z', 
 'visited_hotel')], 
 '19813b0b79ec87975e42e02ff34724dd960c7b05efec71477ec66fb04b6bed9c': [('2018- 
 10-10T18:10:10.242Z', 'referal_code_shared')]
}

我还有一个带有相应列的数据框：

Columns: [visited_hotel, search_hotel, booking_initialized, creferal_code_shared]

我想要做的是遍历每个字典条目，然后适当地将其作为行追加到我的数据框。每行是一个数字，指示用户执行该事件的次数。

例如，在阅读完我的词典摘录后，我的数据框将显示为：

  visited_hotel search_hotel booking_initialized referal_code_shared
0     0             0             1                    0

1     4             1             0                    0

2     0             0             0                    1

预先感谢：）

Answer 1

from collections import Counter
import pandas as pd

# d is your dictionary of values

result = {user: Counter(x[1] for x in records)
          for user, records in d.items()}
df = pd.DataFrame(result).fillna(0).T.reset_index(drop=True)

一种更简洁的方法

result = {i: Counter(x[1] for x in records)
          for i, records in enumerate(d.values()) }

df = pd.DataFrame(result).fillna(0).T

如果要按特定顺序排列列，则

cols = ['visited_hotel', 'search_hotel', 'booking_initialized', 'referal_code_shared']
df = df.loc[:, cols]

Answer 2

d = {
    '56d306892fcf7d8a0563b488bbe72b0df1c42f8b62edf18f68a180eab2ca7dc5': [('2018-10-24T08:30:12.761Z', 'booking_initialized')],
    'ac3406118670ef98ee2e3e76ab0f21edccba7b41fa6e4960eea10d2a4d234845': [('2018-10-20T14:12:35.088Z', 'visited_hotel'), ('2018-10-20T14:17:38.521Z', 'visited_hotel'), ('2018-10-20T14:16:41.968Z', 'visited_hotel'), ('2018-10-20T13:39:36.064Z', 'search_hotel'), ('2018-10-20T13:47:03.086Z', 'visited_hotel')],
    '19813b0b79ec87975e42e02ff34724dd960c7b05efec71477ec66fb04b6bed9c': [('2018-10-10T18:10:10.242Z', 'referal_code_shared')]
}

def user_actions(user, actions):
    # Convert the actions to dataframe
    df = pd.DataFrame(actions).rename(columns={0: 'timestamp', 1: 'action'})
    # Count each action
    counted = df.groupby(['action'])['timestamp'].agg('count').reset_index().rename(columns={'timestamp': 'counter'})
    # Pivot the result so each action is a column
    pivoted = counted.pivot(columns='action', values='counter')
    return pivoted

# Process each user's actions and concatenate all
all_actions_df = pd.concat([user_actions(user, user_actions_list) for user, user_actions_list in d.items()]).replace(np.nan, 0)

输出

    booking_initialized referal_code_shared search_hotel    visited_hotel
0   1.0                 0.0                 0.0                 0.0
0   0.0                 0.0                 1.0                 0.0
1   0.0                 0.0                 0.0                 4.0
0   0.0                 1.0                 0.0                 0.0

如何遍历字典的值并将它们作为行添加到相应的数据框？

2 个答案: