题解 | 牛客网连续练习题目3天及以上的用户

import pandas as pd

# 读取数据集
df = pd.read_csv("nowcoder.csv")
# 将date列转换为日期时间类型
df["date"] = pd.to_datetime(df["date"])
df['datetime']=df['date'].dt.date
df = df[(df["date"].dt.year == 2021) & (df["date"].dt.month == 12)]
df["newrank"] = df["datetime"] - pd.to_timedelta(
    df.groupby("user_id")["datetime"].rank(), unit="d"
)
grouped_df = df.groupby(["user_id", "newrank"]).size().reset_index(name="question_id")
grouped_df=grouped_df[grouped_df['question_id']>=3]
# 按 user_id 分组，取每组中 count 列的最大值
max_count_per_user = grouped_df.groupby("user_id")["question_id"].max()

print(max_count_per_user)