使两列CSV文件用户id显示一次，并以空格分隔会议列表-python

在上面的链接中有很好的解释。但在我的情况是有点不同

user     meetings
178787    287750
178787    151515
178787    158478
576585    896352
576585    985639
576585    456988

预期结果是

user       meetings
178787   "[287750,151515,158478]"
576585   "[896352,985639,456988]"

---

如何使用python和上面的代码实现这一点。提前感谢。

您可以逐行阅读文件，拆分行并将会议添加到字典中，其中关键是用户。使用该方法可以非常巧妙地完成此操作

from collections import defaultdict
import csv

inpath = ''  # Path to input CSV file
outpath = ''  # Path to output CSV file

output = defaultdict(list)  # Dictionary like {user_id: [meetings]}

for row in csv.DictReader(open(inpath)):
    output[row['user']].append(row['meetings'])

with open(outpath, 'w') as f:
    for user, meetings in output.items():
        row = user + ',' + str(meetings) + '\n'
        f.write(row)

然后，我们可以使用制表符将这本词典写回同一个文件，使所有内容对齐

因此，假设您的文件名为f.csv，则代码如下所示：

d = {}
for l in open('f.csv').read().split('\n')[1:-1]:
    u, m = l.split()
    d.setdefault(u, []).append(m)

with open('f.csv', 'w') as f:
    f.write('user\tmeetings\n')
    for u, m in d.items():
        f.write(u + '\t' + str(m) + '\n')

产生以下所需输出：

user    meetings
178787  ['287750', '151515', '158478']
576585  ['896352', '985639', '456988']

既然用户将是关键，我们就编一本字典吧。注意：这最终会将整个文件加载到内存中一次，但不需要先按用户对文件进行排序。还要注意，输出也没有排序，因为dict.items不会以任何确定的顺序检索字典项

output = {}
with f as open('input.csv'):
    for line in f:
        user, meeting = line.strip('\r\n').split()
        # we strip newlines before splitting on whitespace

        if user not in output and user != 'user': 
            # the user was not found in the dict (and we want to skip the header)
            output[user] = [meeting] # add the user, with the first meeting
        else: # user already exists in dict
            output[user].append(meeting) # add meeting to user entry

# print output header
print("user meetings") # I used a single space, feel free to use '\t' etc.
# lets retrieve all meetings per user
for user, meetings in output.items() # in python2, use .iteritems() instead
    meetings = ','.join(_ for _ in meetings) # format ["1","2","3"] to "1,2,3"
    print('{} "[{}]"'.format(user, meetings))

发烧友：排序输出。我首先对键进行排序。注意，这将使用更多的内存，因为我也在创建一个键列表

# same as before
output = {}
with f as open('input.csv'):
for line in f:
    user, meeting = line.strip('\r\n').split()
    # we strip newlines before splitting on whitespace

    if user not in output and user != 'user': 
        # the user was not found in the dict (and we want to skip the header)
        output[user] = [meeting] # add the user, with the first meeting
    else: # user already exists in dict
        output[user].append(meeting) # add meeting to user entry

# print output header
print("user meetings") # I used a single space, feel free to use '\t' etc.

# sort my dict keys before printing them:
for user in sorted(output.keys()):
    meetings = ','.join(_ for _ in output[user])
    print('{} "[{}]"'.format(user, meetings))

熊猫提供了一个很好的解决方案：

import pandas as pd

df = pd.read_csv('myfile.csv', columns=['user', 'meetings'])
df_grouped = df.groupby('user')['meetings'].apply(list).astype(str).reset_index()

---

发布当前代码我喜欢使用dict.setdefaulthere@cowbert是的，这是一个很酷的用法，非常感谢jp_数据分析，它工作得很好。