Subtracting Values Of A List In A Dictionary Using A Dataframe

I have a dataframe below, with the products purchased by users. DataSet : user age maritalstatus product A Young married 111 B young married 222 C young Single 11

Solution 1:

You can use a dict comprehension:

{k:[e for e in v if e not in df.loc[df.user.eq(k), 'product'].tolist()] for k,v in d.items()}
Out[292]: {'A': [222], 'B': [111], 'C': [], 'D': [], 'G': [222], 'X': [222, 444]}

A slightly more verbose solution for easier understanding:

First to build a user-product dict:

user_prod = df.groupby('user')['product'].apply(list).to_dict()
{'A': [111],
 'B': [222],
 'C': [111],
 'D': [222],
 'E': [111],
 'F': [222],
 'G': [555],
 'H': [444],
 'I': [333]}

Then, use a dict comprehension to remove elements which are in the user_prod dict.

{k:[e forein v if e not in user_prod.get(k,[])] fork,v in d.items()}
Out[319]: {'A': [222], 'B': [111], 'C': [], 'D': [], 'G': [222], 'X': [222, 444]}

The use of user_prod.get is necessary because the user may not exist and .get will avoid an exception by returning an empty list.

Solution 2:

Here is one intuitive way to implement your logic. You can optimize via sets and comprehensions, but for reasonable size datasets the below method should be adequate.

products = df.groupby('user')['product'].apply(list)

d = {'A':[111,222], 'B':[111,222], 'C':[111], 'D':[222], 'G':[222,555], 'X':[222,444] } 

fork, v in d.items():
    p = products.get(k)
    if p:
        foriin p:
            d[k].remove(i)

# {'A': [222], 'B': [111], 'C': [], 'D': [], 'G': [222], 'X': [222, 444]}

Solution 3:

   product  user01101211211231134214

new_purchase = frame.set_index('user')['product'].to_dict()

{10:1, 11:2, 12:1, 13:1, 14:2}

{10: [2, 1], 11: [2], 12: [], 13: [22], 14: [1]}

result = {}
        fork, v in prev_purchase.items():
            result[k] = [item foritemin v if item not in [new_purchase[k]]]

{10: [2], 11: [], 12: [], 13: [22], 14: [1]}