Update iceberg-spark to use Hive tables

[rdblue]

The Spark integration currently uses path-based tables and the HadoopTables implementation. Now that iceberg-hive is available, Spark should support named tables loaded with HiveTables.

[aokolnychyi]

@rdblue How do you see it in 2.4? I understand that HiveTables will be used when no path is given but how will we fetch the table identifier? As another data source property?

Also, how will it look like when we have support for CREATE TABLE statements, where LOCATION says if it is a managed or external table?

[rdblue]

I was planning on adding support by inspecting the path. If it contains /, then it is a path and should use HadoopTables. If it doesn't, then use it as a table identifier with HiveTables.

[aokolnychyi]

@rdblue Have you already started working on this? If not, we can contribute this part if it doesn't depend on the catalog effort.