Python 如何分组并查找新的或正在消失的项目

你可以把它作为一个set:

offers = df.groupby('date', sort=True)['offer-id'].agg(set)

date
2024-02-10       {1, 2, 5}
2024-02-11    {1, 2, 3, 4}
Name: offer-id, dtype: object

然后，拿到diff美元，你就会得到新的物品:

offers.diff()

date
2024-02-10       NaN
2024-02-11    {3, 4}
Name: offer-id, dtype: object

或已售出的物品:

offers.diff(-1)

date
2024-02-10    {5}
2024-02-11    NaN
Name: offer-id, dtype: object

如果你想要物品的数量，链str.len:

offers.diff().str.len().fillna(0).convert_dtypes()

date
2024-02-10    0
2024-02-11    2
Name: offer-id, dtype: Int64

要获得这些作为新专栏的信息，有map个:

df['new'] = df['date'].map(offers.diff().str.len().fillna(0).convert_dtypes())
df['sold'] = df['date'].map(offers.diff(-1).str.len().fillna(0).convert_dtypes())

print(df)

输出:

   offer-id        date  price  new  sold
0         1  2024-02-10     30    0     1
1         1  2024-02-11     10    2     0
2         2  2024-02-10     30    0     1
3         2  2024-02-11     30    2     0
4         3  2024-02-11     20    2     0
5         4  2024-02-11     25    2     0
6         5  2024-02-10     20    0     1