[WIP] copy on write plus merge

[dilipbiswal]

• Adds WIP support of MERGE INTO leveraging the work for DELETE
• Implements copy-on-write

Plan:

== Optimized Logical Plan ==
ReplaceData RelationV2[key1#50, value1#51] file:///..., IcebergWrite(table=file:///..., format=PARQUET)
+- MergeInto org.apache.spark.sql.catalyst.plans.logical.MergeIntoProcessor@e1a150c, RelationV2[key1#50, value1#51] file:///...
   +- Join FullOuter, (key1#50 = key2#65)
      :- Project [key2#65, value2#66, true AS _source_row_present_#138]
      :  +- RelationV2[key2#65, value2#66] file:///...
      +- Project [key1#50, value1#51, true AS _target_row_present_#139]
         +- DynamicFileFilter
            :- RelationV2[key1#50, value1#51] file:///...
            +- Aggregate [_file_name_#137], [_file_name_#137]
               +- Project [_file_name_#137]
                  +- Join Inner, (key1#50 = key2#65)
                     :- Filter isnotnull(key2#65)
                     :  +- RelationV2[key2#65] file:///...
                     +- Project [key1#50, input_file_name() AS _file_name_#137]
                        +- Filter isnotnull(key1#50)
                           +- RelationV2[key1#50] file:///...