Re: Sequence ID from INSERT

Daryl,

your second example where you get the id first is the only way I know
how to do it

BTW it doesn't have to be done in a transaction. the id will not be used
by another connection.

DAve
On Tue, 2002-07-23 at 11:05, Daryl Beattie wrote:
> Dear Postgresql-JDBCers,
>
>     Let's say I have created a table like this:
>
> CREATE TABLE Food (
>     Id    SERIAL PRIMARY KEY,
>     Name    VARCHAR(10)
> );
>
>     What is the best way to use JDBC to insert a row into the database
> and get the Sequence Id right out of it? I don't want to start a transaction
> and have to do another query that looks like this:
>
> BEGIN;
> INSERT INTO Food (Name) VALUES ('Apple');
> SELECT currval('Food_Id_seq');
> COMMIT;
>
>     Or ones like these:
>
> BEGIN;
> new_id = output of "SELECT nextval('Food_Id_seq')";
> INSERT INTO Food (Id, Name) VALUES (new_id, 'Apple');
> COMMIT;
>
> BEGIN;
> INSERT INTO Food (Name) VALUES ('Apple');
> SELECT MAX(Id) FROM Food; -- will check ResultSet of this separate query.
> COMMIT;
>
>     I want to do it in one statement. Normally I would
> setAutoCommit(false), execute two separate statements, and then commit().
> However, I was wondering if there's a way to do all this in a ... "cleaner"
> fashion? Can anybody provide me with a sample?
>
> Thanks,
>
>     Daryl.
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org
>
>