error when train model without LabelEncoder or using LabelEncoder fit with lager data and transform with lower data.

[Nhanvu1009]

i've problem with LabelEndcoder:

First: I'm training model with movielens data set. It have ["userId", "movieId", "year", "genres", "overview", "tagline"] attribute. in this, the userId, movieId, year is LINEAR_FEATURE_COLUMNS. in paper i see the author using sklearn.labelencoder to LINEAR_FEATURE_COLUMNS. i was try not using sklearn.labelencoder but it have an error.

but i compare the type, size, value count,... of original feature and labelencoder feature are same.

Second: i'm using labelencoder with lager feature but i use part of them to train model, it have the same error.

anyone have same problem? tell me how to solve this