SQL & PL/SQL

Is it possible to create a procedure that does a merge (upsert) having only two parameters:

1. Existing Table Name - name of table that will have data merged into it.
2. Incoming Table Name - name of the table that has updated/new data.

The methodology would be:

1. Uses the primary key(s) on the "existing table" as the linking condition
2. Does and update if exists and insert if new
3. no logging
4. no hints
5. no delete clause

Oracle 11g

MERGE <hint> INTO <table_name>

USING <table_view_or_query>

ON (<condition>)

WHEN MATCHED THEN <update_clause>

DELETE <where_clause>

WHEN NOT MATCHED THEN <insert_clause>

[LOG ERRORS <log_errors_clause> <reject limit <integer | unlimited>];

<table_name> - taken from input parameter

<table_view_or_query> - taken from input parameter

<condition> - links primary key (how do we look up a single/multiple primary keys and return the list?)

<update_clause> - update all columns (other than the PK, how to get and use a dynamic list of columns)

<where_clause> - not used

<insert_clause> - insert new records for all columns (including the PK(s))