Is there a way to get my output organized by unique user ID?

I have a massive dataset (practice Instacart data) I'm working on in Python pandas. I'm trying to create customer profiles, but have just realized that the set includes over 32 million rows(which are unique orders), but ONLY 206,000 unique customers.

user_id = [1, 2, 1, 1, 5, 3, 3, 2, 4, 6]
department = [produce, pets, frozen, pets, pets, meat, other, pets, snacks, snacks]
order_id = [1, 1, 2, 3, 1, 1, 2, 2, 1, 1]
income = [72000, 61888, 72000, 72000, 42867, 85629, 85629, 61888, 57211, 53665]

     user_id department order_id income
0    1       produce    1        72000
1    2       pets       1        61888
2    1       frozen     2        72000
3    1       pets       3        72000 
4    5       pets       1        42867
5    3       meat       1        85629
6    3       other      2        85629
7    2       pets       2        61888
8    4       snacks     1        57211
9    6       snacks     1        53665

For example, I'd like to create a profile for pet owners by looking for entries that indicate 'pet' in the department column. The issue is that when I count the number of 'pet' entries in the 'department' column, I get 4. The reality is that there are only 3 different users that bought things for pets, not 4.

In order to create an accurate customer profile, I need to be able to know how many unique users shopped in different departments whereas the dataset as it is counts instances.
My apologies if this is not clear, it's my first post for help!

I started with this:

df.loc[df['department'] == 'pets', 'pet_owner'] = 'Yes'
df.loc[df['department'] != 'pets', 'pet_owner'] = 'No'

Then I realized the analysis issue with just looking at how many ORDERS included pet items, rather than how many unique customers made an order including pet items.

Then I tried this, which was recommended by another user:

owns_pets = list(df[df["department"] == 'pets']['user_id'].unique())

no_pets = list(df[df["department"] != 'pets']['user_id'].unique())    

owns_pets.value_counts(dropna = False)

Out: AttributeError: 'list' object has no attribute 'value_counts'

So I changed tactics again and tried the following:

df.groupby('user_id').agg({'pet_owner' : 'count'})

Out: user_id  pet_owner
     	
      1	           59
      10	       143
      100	       27
      1000	       103
      10000	       1092
      ...	       ...
      99995	       50
      99996	       128
      99997	       36  

Clearly the count is returning each instance rather than just the result 'per' unique user_id.

Can anyone please help me figure out how to get the info the way in which I need? TIA

While it is difficult to be sure of exactly what you want, since you haven't provided any sample data nor provided a good example of the desired output. Here is an approach to select the unique Cust_ID's that have made purchases involving the Pet department.

Given a dF of the following form:

	Cust_ID	Purch_Date	Dept
0	4817	2022-05-26	Pets
1	3013	2022-01-12	Pets
2	3013	2022-12-22	Hardware
3	4550	2022-04-21	Pets
4	4817	2022-12-26	Hardware

You can get a list of the unique cust_id's as follows:

list(df[df["Dept"] == 'Pets']['Cust_ID'].unique())  

This will produce a list of just the cust_ids satisfying the condition that Dept contains 'Pets' as shown below:]

[4817, 3013, 4550]

If all you need is the number of unique users, just take your answer as:

len(list(df[df["Dept"] == 'Pets']['Cust_ID'].unique()))

Which in my sample case yields 3

I apologize for being unclear with my question. I did end up getting it answered on another forum though and here is the solution:

df10exuser = df.drop_duplicates(subset = ['user_id'])