CODEWITHSUNDEEP
pythonpandasdataframegroup-by
techytushar
techytushar

Reputation: 803

Select users with specific number of orders pandas

I have a dataframe with user_id and order_number columns. order_number tells the nth order by a user. I want to select users who have done certain number of orders.

Sample DataFrame:

user_id order_number
0   1   1
1   1   2
2   1   3
3   2   1
4   2   2
5   3   1
6   3   2
7   3   3

Output: [1,3] Output should be user_id [1,3] because they have done 3 orders but 2 have done only 2 orders.

I am trying:

(df.groupby(['user_id'])['order_number'].max()==3)

This gives me boolean series but how to select index with only True values?

Upvotes: 1

Views: 298

Answers (1)

anky
anky

Reputation: 75080

A general way of doing this is by using df.loc[] or df.query:

df.groupby(['user_id'],as_index=False)['order_number'].max().query("order_number==3")
#or
df.groupby(['user_id'],as_index=False)['order_number'].max().loc[
                                   lambda x: x['order_number']==3]

For this example you dont have to get max of another column, you can just count them, as @Steven suggests :

df.groupby('user_id',as_index=False).count().query("order_number==3")

Or as @Wen suggests:

df['user_id'].value_counts().loc[lambda x: x==3]

   user_id  order_number
0        1             3
2        3             3

Upvotes: 6

Related Questions