Should we extract the RootRel names using the Iceberg table or should we query it from the SQL query received. In theory, we would get a Substrait plan which contains the evolved schema.
The fields in the projection output (if it is the final relation), should show the number of columns that we expect in the final output.
That number of columns and the number of columns in the iceberg table must be same as long as we use the selected_fields parameter (populated by fields in the SQL query at the moment) when creating the Iceberg scanner.
Should we extract the
RootRelnames using the Iceberg table or should we query it from the SQL query received. In theory, we would get a Substrait plan which contains the evolved schema.The fields in the projection output (if it is the final relation), should show the number of columns that we expect in the final output. That number of columns and the number of columns in the iceberg table must be same as long as we use the
selected_fieldsparameter (populated by fields in the SQL query at the moment) when creating the Iceberg scanner.We should validate this logic.