Exclude a few columns from source table in the TO TRAIN | PREDICT | EVALUATE statement.

[brightcoder01]

Given a wide table containing many columns (such as 100 or more), there are a few columns (such as 5) which are neither feature nor label for the training statement. In the SQLFlow TO TRAIN statement, we don't want to feed the unwanted columns into the model. How do we exclude them in our SQL statement?

Currently, we can write the names of the necessary columns (100 - 5 = 95 columns) specifically in the 'SELECT ... FROM' clause. But it will make the SQL statement too long. We need a solution to make our SQL statement simple and short.