Re: Oracle 8 and objects ?
Date: 31 Jan 94 03:16:33 GMT
Message-ID: <Jan.30.22.16.32.1994.15517_at_andromeda.rutgers.edu>
perera_at_pollux.cs.uga.edu (perera) writes:
>Hi:
>I understand that the next version of Oracle will support "Objects". How
>does one get a table to inherit another table ? Does anybody have any
>insight on this ( syntax wise ) ? All suggestion are wellcome.
The syntax I'm not sure of. Most likely the SAG will come up with some type of syntax in the near future.
For now, there are basically two approaches to supporting "objects" and object oriented concepts. The first is to place some type of object interface or "object wrapper" as it is sometimes called, around a relational model. What this will do is give an underlying relational storage system the appearance of objects.
For example, a table's data definition (attributes and data types) may be loosely considered the data portion of an class. Triggers which are definied to work against the table and, in fact, provide the only means of accessing data in the table, might be similar to methods. The encapsulation point should be clear in this case. Only those triggers (methods) defined with a table should be allowed to access the data.
Inheritance can then be seen as taking the data definition and triggers of a table and using them as a starting point for a new table which may add new data definitions and/or triggers.
One possible way to do this might be to simply use a pre-defined table (class) as a data type when defining a new table.
The last piece, polymorphism, is much more difficult to approximate. Oracle 7 does have the DBMS_PIPE type function but one needs to know the senders and receivers before hand.
Mind you, any OO researcher would laugh his or her head off at the liberties I've taken but the analogies are clear.
BTW, the second way to go about it is to take an object oriented database and provide support for relational concepts. This can be done by a few comercial ODBMSs.
Rich Holowczak
Rutgers University
holowcza_at_andromeda.rutgers.edu
Received on Mon Jan 31 1994 - 04:16:33 CET