Please post DDL, so that people do not have to guess what the keys,
constraints, Declarative Referential Integrity, datatypes, etc. in
your schema are. Sample data is also a good ideas, along with clear
specifications.
1) Stop putting those silly, redundant "t_" prefixes to tell us that a
table is a table. I am sorry that business is so bad that you have
only one account instead of a talbe to model Accounts (note the use of
plurals or collective nouns for sets). You need to read the ISO-11179
standards for data elements name
Quote:
I have an t_account table with acc_id (identity type) as PK |
An IDENTITY property can, by definition, never be a key. And it means
that you have no checkdigits, no verification in the real world that
your model is correct.
Quote:
I have another table t_customer with cust_id (identity type) as PK. |
Ditto.
Quote:
t_account and t_customer share a many-to-many relationship. So I
have a third table t_acc_cust to reflect the many to many
|
relationship. <<
No DDL, so I can only hope that you have something like this:
CREATE TABLE CustAccts
(acct_nbr INTEGER NOT NULL
REFERENCES Accounts(acct_nbr)
ON UPDATE CASCADE
ON DELETE CASCADE,
cust_nbr INTEGER NOT NULL
REFERENCES Customers(cust_nbr)
ON UPDATE CASCADE
ON DELETE CASCADE,
PRIMARY KEY(acct_nbr, cust_nbr));
But I have the feeling that anyone who would use an IDENTITY for a key
would not bother with DRI and might even have put an IDENTITY on this
table as a pseudo-key.
Quote:
How can I write the SQL statements so the the t_account, t_customer
tables get inserted with new values and also the t_acc_cust table gets
|
populated with acc_id and cust_id. <<
You need three INSERT INTO statements. The account and the customer
must exist in their tables before you can have a relationship between
them. This will be easy once you have a proper design.