Programmatically create insert statements for Oracle and SQL Server

Building the insert statements, I want to pass a table name and where clause and have it return all of the insert statements neccessary to recreate those records in another database which has that same table.

well, you know the INSERT syntax:

INSERT INTO destination_table_name
  ( col1, col2, col3, col4 .... et c. ... )
VALUES
  ( value1, value2, value3, value4 .... etc ... )

so, I repeat: where is the problem.
you know the destination table.
you know the destination column names
you queried the source values
the only things you must watch out, for the values in the sql insert, is
* NULL  
VALUES ( ... , NULL ,  ... )
* numerical values (decimals)
VALUES ( ... , 120000.23434  ,  ... )
* strings (single quotes around, and 2 single quotes for a string value containing single quotes
VALUES ( ... , 'Mc''Donalds',  ... )
* dates (implicit data type conversion)
Oracle:
VALUES ( ... , TO_DATE('2009-08-01', 'YYYY-MM-DD') , ...
SQL Server:
VALUES ( ... , CONVERT(datetime, '2009-08-01', 120) , ...

hope this helps

Ok i will try and rephrase again, I won't know the column names, number of columns or the type of each column because like i said I want the actual table to be a parameter that means I want to use it on many tables on demand as I need in my .Net application.  Of course I can hard code for each table but that is precisely what I am trying to avoid.

will the data type in the source table be effectively the same as in the destination table?

if yes, you can of course know the column names and data types, by querying the catalog of the table.
in oracle, check the COL view:
http://ss64.com/orad/COL.html

The data types will be the exact same in the Oracle destination and effectively the same in the SQL Server destination.  I know I can build the routine to do this by querying the database catalog I was just hoping someone had already done the grunt work and had an available Stored Procedure or code block to handle it.  It seems that I am going to have to bite the bullet and write it myself.

Thanks for your help, I actually did find a procedure that works when using SQL server as it's source but it is far from compatible in oracle.

Yeah I had found that before I posted but could never get it to compile or run in Toad

what version of oracle?
what error(s) ?

Oracle 10G, Using TOAD as my SQL Editor

ORA-06502: PL/SQL: numeric or value error: character string buffer too small at line 62.

It does work on a table with a smaller amount of fields, but all of the tables I need to use are blowing up...

I see. the VARCHAR2(4000) is not big enough :)

I solved that, in one of my other developments, with a function that returned data in a "pipelined" manner, aka like a rowset (but not with a cursor):
http://www.akadia.com/services/ora_pipe_functions.html

for your, you could return 1 table value per "returned row", for example...