将值从一个csv文件匹配到另一个，并使用pandas / python替换整个列

Question

Consider the following example:

I have a dataset of Movielens-

u.item.csv

ID|MOVIE NAME (YEAR)|REL.DATE|NULL|IMDB LINK|A|B|C|D|E|F|G|H|I|J|K|L|M|N|O|P|Q|R|S|
1|Toy Story (1995)|01-Jan-1995||http://us.imdb.com/M/title-exact?Toy%20Story%20(1995)|0|0|0|1|1|1|0|0|0|0|0|0|0|0|0|0|0|0|0
2|GoldenEye (1995)|01-Jan-1995||http://us.imdb.com/M/title-exact?GoldenEye%20(1995)|0|1|1|0|0|0|0|0|0|0|0|0|0|0|0|0|1|0|0
3|Four Rooms (1995)|01-Jan-1995||http://us.imdb.com/M/title-exact?Four%20Rooms%20(1995)|0|0|0|0|0|0|0|0|0|0|0|0|0|0|0|0|1|0|0

Seperator used here is Pipe, which is still manageable.

training_data.csv

,user_id,movie_id,rating,unix_timestamp
0,1,1,5,874965758
1,1,2,3,876893171
2,1,3,4,878542960

Since I need to show the Movie names in "Training_data", instead of "movie id" I need to match every ID of u.item.csv with movie_id with training_data.csv and then replace it.

I'm using Python Pandas, and The training data was converted from Sframe to Dataframe to CSV. So that I could acquire the required change, which is yet unsuccessful. I can surely use some looping structure, but matching and replacing is real challenge I face.

Ps I know Training data will be in sequence per user and will produce the exact output if replaced, but I need to learn this so that when I recommend movies I need MOVIE Names to displayed and not IDs.

I've already tried

1. THIS (pandas-python-replace-multiple-values-in-multiple-columns) - But can cost a lot of time when I have 100K values in Dataset
2. THIS (pandas-replace-multiple-values-one-column) - Matching values not explained
3. THIS (pandas-replacing-column-values) - Manual entries are done

Answer 1

I think you need map by Series created by set_index :

print (df1.set_index('ID')['MOVIE NAME (YEAR)'])
ID
1     Toy Story (1995)
2     GoldenEye (1995)
3    Four Rooms (1995)
Name: MOVIE NAME (YEAR), dtype: object

df2['movie_id'] = df2['movie_id'].map(df1.set_index('ID')['MOVIE NAME (YEAR)'])
print (df2)
   user_id           movie_id  rating  unix_timestamp
0        1   Toy Story (1995)       5       874965758
1        1   GoldenEye (1995)       3       876893171
2        1  Four Rooms (1995)       4       878542960

Or use replace :

df2['movie_id'] = df2['movie_id'].replace(df1.set_index('ID')['MOVIE NAME (YEAR)'])
print (df2)
   user_id           movie_id  rating  unix_timestamp
0        1   Toy Story (1995)       5       874965758
1        1   GoldenEye (1995)       3       876893171
2        1  Four Rooms (1995)       4       878542960

Difference is if not match, map create NaN and replace let original value:

print (df2)
   user_id  movie_id  rating  unix_timestamp
0        1         1       5       874965758
1        1         2       3       876893171
2        1         5       4       878542960 <- 5 not match

df2['movie_id'] = df2['movie_id'].map(df1.set_index('ID')['MOVIE NAME (YEAR)'])
print (df2)
   user_id          movie_id  rating  unix_timestamp
0        1  Toy Story (1995)       5       874965758
1        1  GoldenEye (1995)       3       876893171
2        1               NaN       4       878542960

---

df2['movie_id'] = df2['movie_id'].replace(df1.set_index('ID')['MOVIE NAME (YEAR)'])
print (df2)
   user_id          movie_id  rating  unix_timestamp
0        1  Toy Story (1995)       5       874965758
1        1  GoldenEye (1995)       3       876893171
2        1                 5       4       878542960