Modifying INSERT
From: Laconic2 <laconic2_at_comcast.net>
Date: Tue, 6 Apr 2004 10:23:05 -0400
Message-ID: <R4KdncfoF_2CIO_dRVn-sA_at_comcast.com>
Let's say that a USER (programmed or human) does an INSERT of a new table row that duplicates, column for column, the values in an existing row.
Date: Tue, 6 Apr 2004 10:23:05 -0400
Message-ID: <R4KdncfoF_2CIO_dRVn-sA_at_comcast.com>
Let's say that a USER (programmed or human) does an INSERT of a new table row that duplicates, column for column, the values in an existing row.
Present day DBMS products act differently, depending on whether the table has a primary key or not.
If there is no primary key, the INSERT creates a duplicate row.
If there is a primary key, the insert fails.
But how about a different behaviour? What if the DBMS reported success back to the writer, but didn't write anything to the table? Essentially, the new value of the table is the UNION of the existing table and the inserted row?
What would the consequences be of this behaviour?
Received on Tue Apr 06 2004 - 16:23:05 CEST