Hi, I have a dataset which contains numerical columns, categorical columns and SQL Query as a separate column. Can someone recommend me how to deal with it?
I would like to get suggestion on how to deal with SQL Query column? what sort of feature engineering should I apply for SQL Query column?
Is there any specific model to deal with this?
This really depends on what you’re trying to achieve. If you’re doing a pure numerical prediction for example, the SQL query may not add any value to your model.
On the other hand, if you’re looking at generating SQL through, starting off with a standard NLP approach of tokenising, then developing a language model would be a good start.
@JamesTrick really?
I wanted to predict numerical value, and I feel SQL query will add weightage to the model. so, hence why I thought to check with community, is there any better feature engineering method to deal with SQL query.