How can I fill empty ID cells with the IDs from either one of two other DataFrames

Question

I have the following table in which some (but not all) user IDs are missing:

user ID	item ID	item type
10	123	question
NaN	126	answer
14	129	question

To get the missing user IDs I want to look up the corresponding item ID in the following two tables (depending on whether the item type is a question or an answer in the table above).

answer DataFrame:

item ID	user ID
126	12

question DataFrame:

item ID	user ID
123	10
129	14

Finally, I want to get something like this:

user ID	item ID	item type
10	123	question
12	126	answer
14	129	question

Answer 1

Try with numpy.select:

import numpy as np

conditions = [df["user ID"].isnull() & df["item type"].eq("question"),
              df["user ID"].isnull() & df["item type"].eq("answer")]

choices = [df["item ID"].map(dict(zip(question["item ID"],question["user ID"]))),
           df["item ID"].map(dict(zip(answer["item ID"],question["user ID"])))]

df["user ID"] = np.select(conditions, choices, df["user ID"])

>>> df
   user ID  item ID item type
0     10.0      123  question
1     10.0      126    answer
2     14.0      129  question

Answer 2

You can use a np.where() and a merge to get the data you need

df['user ID'] = df['user ID'].fillna(0).astype(int)

df_final = pd.merge(left = df, right = answer_df, on = 'item ID', how = 'outer', suffixes = ('', '_right'))
df_final['user ID'] = np.where(df_final['user ID'] == 0, df_final['user ID_right'], df_final['user ID']).astype(int)
df_final[['user ID', 'item ID', 'item type']]