How to cast a DataFrame to a dictionary/dataframe of the right kind using Pandas?

Question

Python

There is a select query.

print("Connecting to database...") 
db = psycopg2.connect(host='localhost', user='postgres', password = 'password', database = 'DB') 
print("Connected!") 
print("Processing...") 
prefs = pd.read_sql("select distinct ord.user_id, line.variant_id, count(line.order_id) " + 
"from spree_line_items line " + 
"join spree_orders ord on(line.order_id=ord.id) " + 
"where ord.state='shipped' and ord.user_id is not null " + 
"group by line.variant_id, ord.user_id " + 
"order by ord.user_id;" 
, con=db)
prefs.columns=['user', 'var', 'mark']
print("Adding marks...") 
prefs.loc[prefs.mark>1, 'mark']=5
prefs.loc[prefs.mark==1, 'mark']=4

Ratings are given if the user has ordered product variant 1 or more times with ratings 4 and 5, respectively.
When outputting the resulting dataframe, I get the following picture:

Attention, question! How to make a dictionary of the form from a given dataframe: df_dict={user1:{var_id1 : mark1, var_id2 : mark2...},..., usern: {var_id : mark,....}} or arrange a dataframe where user is 1st row, and var is the indexing column, then mark is the cell of the user matrix on var.

Answer 1

pandas.pydata.org/pandas-docs/stable/generated/pan...