训练集按照feat去重

tensnsorflow/ffm.py

@@ -137,6 +137,7 @@ class multiFFMFormatPandas:
             return False
 
 def check():
+    print("check")
     db = pymysql.connect(host='10.66.157.22', port=4000, user='root', passwd='3SYz54LS9#^9sBvC', db='jerry_test')
     sql = "select max(stat_date) from esmm_train_data"
     validate_date = con_sql(db, sql)[0].values.tolist()[0]
@@ -173,7 +174,7 @@ def check():
     print("before drop duplicate")
     print(df.shape[0])
     print("after drop duplicate")
-    df = df.drop_duplicates()
+    df = df.drop_duplicates("feat")
     print(df.shape[0])
     print("after group by")
     print(len(df.groupby("feat")))