I want to describe parquet files via views so that they are visible to the JDBC driver (i.e. DatabaseMetadata.getTables).
Further, I want to define a column type (i.e. integer) versus leaving it to default to the ANY type.
One cannot CAST(<exp> as <complex type>) such as ARRAY<INTEGER> or MAP.
Can you manually create the view.drill JSON file where the columns are not projected with CAST and the fields JSON array includes the type ARRAY etc?
Is the view.drill JSON syntax documented anywhere?