I tried with DB (MySQL) sequence in EJB2.0. I have the following problems.
The column/field (orderID) using the sequence is primary key. My code is
like this:
public abstract class OrderRequestBean implements EntityBean {
//........
public java.lang.Integer ejbCreate(java.lang.String userid, String
orderInfo) throws CreateException {
setOrderID(new Integer(0));
/*
orderID uses sequence feature in MySQL. But as the primary key, it has to be
set a non-null
value in ejbCreate method. This is the 1st problem.
^^^^^^^^^^^^^^^^
For MySQL, you can set 0, and it will generate the next sequence
number and replace 0.
*/
setUserid(userid);
setOrderInfo(orderInfo);
return null;
}
//.....
}
public class OrderControllerBean implements SessionBean {
//..............
public int processOrder(Order order) {
//..............
try {
Object ref = ctx.lookup("OrderRequest");
String user = "alvin";
OrderRequestHome orderRequestHome =
(OrderRequestHome )PortableRemoteObject.narrow(ref, OrderRequestHome
.class);
OrderRequest orderRequestHome = rqHome.create(user, starttime,
orderInfo);
return orderRequest.getOrderID().intValue();
/*
It will return 0!! Not the real sequence number created in DB. So how can I
get the
orderID and return to the client? This is the 2nd problem.
^^^^^^^^^^^^^^^^^
*/
} catch (Exception e) {
e.printStackTrace();
}
}
//..................
}
Any comments?
> Delivered-To: [EMAIL PROTECTED]
> Date: Tue, 19 Mar 2002 10:33:45 -0800
> Reply-To: Gene Chuang <[EMAIL PROTECTED]>
> Sender: A mailing list for Enterprise JavaBeans development
> <[EMAIL PROTECTED]>
> From: Gene Chuang <[EMAIL PROTECTED]>
> Subject: Re: High performance synchronized counter?
> To: [EMAIL PROTECTED]
>
>
> Try db sequence? You'll find your counter worthless the moment you
upgrade
> your
> enterprise app to a clustered environment.
>
>
> Gene
> --- Alvin Wang <[EMAIL PROTECTED]> wrote:
> > Hi! Gurus,
> >
> > I implemented a synchronized counter (for example Order ID) using Entity
> > bean and use exclusive concurrency control perperty in WL to ganrantee
that
> > only one transacation is reading/writing that bean at one time point, so
> > that the generated order id will not duplicate. However, I found it is a
> > bottleneck. So what is the better design to improve the performance? Or
is
> > there other way to generate non-duplicate ID strings?
> >
> > Thanks!
>
>
>
>
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff EJB-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".
