I am using Apache Calcite to develop my software.My target is to simply provide an SQL string to Calcite and mainly using its optimizer to optimize it and output a optimized SQL string, which means that i don't have concrete tables or fields. But as far as I know for now, after the parse process, I must go through a validation process and then I can turn validated SqlNodes to RelNodes.
As the problem above said. Are there any methods I can use to avoid validation process and directly turn the parsed SqlNode to a RelNode?
Validation is required because it's very difficult to provide any optimization without any knowledge of the underlying schema. You may want to look at Apache Drill for a solution.