Input
- ItemTable: cfilter_sports:
col1_item1 col1_item2 cntb ---------- ---------- ---- tennis soccer 2 soccer soccer 2 golf baseball 1 tennis soccer 2 soccer soccer 2 golf baseball 1 soccer tennis 2 basketball baseball 1 soccer tennis 2 basketball baseball 1 basketball baseball 1 basketball baseball 1 basketball baseball 1 basketball baseball 1 baseball basketball 1 baseball basketball 1 baseball basketball 1 baseball basketball 1 baseball basketball 1 baseball basketball 1 baseball baseball 1 baseball golf 1 baseball baseball 1 baseball golf 1 baseball baseball 1 baseball baseball 1 baseball baseball 1 baseball baseball 1 baseball baseball 1 baseball baseball 1
(This table is the same as cfilter_sports_out, output by CFilter Example.)
- UserTable: new_players:
sport new_player preference ---------- ---------- ---------- lacrosse player_e 8 soccer player_c 6 golf player_d 6 tennis player_d 5 basketball player_a 9 basketball player_b 3 baseball player_a 1 baseball player_b 2
SQL Call
SELECT * FROM WSRecommender (
ON (
SELECT * FROM WSRecommenderReduce (
ON cfilter_sports AS ItemTable PARTITION BY col1_item1
ON new_players AS UserTable PARTITION BY sport
USING
Item1 ('col1_item1')
Item2 ('col1_item2')
ItemSimilarityScore ('cntb')
UserIDColumn ('new_player')
UserItemColumn ('sport')
UserPrefColumn ('preference')
) AS dt1
) PARTITION BY usr, col1_item2
) AS dt ORDER BY recommendation DESC, item;
Output
SELECT * FROM cfilter_sports_out ORDER BY 1,2,3;
item usr recommendation new_reco_flag ---------- -------- ------------------ ------------- baseball player_d 6.0 1 soccer player_c 6.0 0 tennis player_c 6.0 1 soccer player_d 5.0 1 baseball player_a 4.4285712242126465 0 baseball player_b 2.4285714626312256 0 basketball player_b 2.0 0 golf player_b 2.0 1 basketball player_a 1.0 0 golf player_a 1.0 1 golf player_d 0.0 0 tennis player_d 0.0 0