Re: prepared statements and sequences

Ryan,

Scroll Way down :)


On Wed, 2003-04-02 at 14:10, Ryan Wexler wrote:
> Dave I couple more comments on the sequence
>
>
> On 2 Apr 2003, Dave Cramer wrote:
>
> > See below, comments on sequence, I will have to look at the date problem
> >
> > On Wed, 2003-04-02 at 10:52, Ryan Wexler wrote:
> > > Thanks for replying much appreicated my comments are inserted
> > >
> > >
> > > On 2 Apr 2003, Dave Cramer wrote:
> > >
> > > >
> > > > Ryan,
> > > >
> > > > See my comments below
> > > >
> > > > On Wed, 2003-04-02 at 00:43, Ryan Wexler wrote:
> > > > > I am interfacing a postgresql db with jdbc using jdk1.4x on a linux box.
> > > > > I have two questions/problems that I need help with.
> > > > >
> > > > > 1)Prepared Statments
> > > > > I am trying to use a prepared statement and am successful except for date
> > > > > fields.  I get a parse error when using the preparedStatement.setDate(x,
> > > > > java.sql.Date);  Is this a postgres thing or a personal problem?  Is there
> > > > > a workaround?
> > > > >
> > > > Can you reproduce this in a small file? There was a similar question
> > > > yesterday??
> > > > >
> > >
> > > Here is the method I am calling.  It is throwing the error on:
> > > pStatement.setDate(4, new java.sql.Date(new java.util.Date().getTime()))
> > >
> > > The connection broker i created uses the "org.postgresql.Driver" driver
> > > and has autocommit set to true.
> > >
> > >
> > >  public void insertOrder()
> > >     {
> > >     try
> > >       {
> > >       String query = "insert into customerorder (customer_id, address_id,
> > > payment_id, createdate, ordertotal, tax, shipping, subtotal) " +
> > >        "values (?, ?, ?, ?, ?, ?, ?, ?)" ;
> > >        System.err.println(query);
> > >
> > >       PreparedStatement pStatement= postgres.getPreparedStatement(query);
> > > pStatement.setInt(1, 1);
> > > pStatement.setInt(2, 1);
> > > pStatement.setInt(3, 1);
> > > pStatement.setDate(4, new java.sql.Date(new java.util.Date().getTime()));
> > > pStatement.setBigDecimal(5, (new BigDecimal(123)).setScale(2,
> > > BigDecimal.ROUND_HALF_UP));
> > > pStatement.setBigDecimal(6, (new BigDecimal(123)).setScale(2,
> > > BigDecimal.ROUND_HALF_UP));
> > > pStatement.setBigDecimal(7, (new BigDecimal(123)).setScale(2,
> > > BigDecimal.ROUND_HALF_UP));
> > > pStatement.setBigDecimal(8, (new BigDecimal(123)).setScale(2,
> > > BigDecimal.ROUND_HALF_UP));
> > >       int i = pStatement.executeUpdate(query);
> > >       System.err.println("i: " + i);
> > >
> > >       }
> > >     catch (Exception x)
> > >       {
> > >       System.err.println("Exception: " + x);
> > >       x.printStackTrace();
> > >       }
> > >
> > >     }
> > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > 2)Sequences-
> > > > > I am using sequences as unique identifiers, or rather I should say I would
> > > > > like to use sequences.  I have successfully set up several sequences and
> > > > > every time i insert an new row it automatically increments itself.  My
> > > > > problem is whenever I insert a row I need to know
> > > > > what the sequence is that was associated with the row inserted.  I
> > > > > can't rely on doing a
> > > > > max(sequenceid) kind of query because there maybe 10 rows inserted in that
> > > > > time.  My method of inserting rows is just using a prepared statement and
> > > > > in my insert statement i don't reference the sequence.  Is there a way to
> > > > > get it to return the sequence id say when you call executeUpdate() on the
> > > > > prepared statement?  Or what is the proper way to do this?
> > > >
> > > > There is no way to get it to return the sequence. However you have two
> > > > options here
> > > >
> > > > 1) get the sequence before the insert and insert it with the data.
> > > >
> > > > select nextval('sequence')
> > > >
> > > > 2) get the sequence after the insert
> > > >
> > > > select currval('sequence')
> > > >
> > > > Both of these methods are multi-connection safe, in other words if two
> > > > connections are creating sequences at the same time, you will get the
> > > > right data.
> > > >
> > >
> > > ???
> > > 1)If you use the nextval('sequence') method then can I assume when you
> > > create your table you don't auto set it to be connected to the sequence
> > > like:
> > > CREATE TABLE "customerorder"
> > >   (
> > >   order_id integer DEFAULT nextval('order_id') UNIQUE not null,
> > >   customer_id integer not null,
> > >   address_id integer not null,
> > >   payment_id integer not null,
> > >   createdate date not null,
> > >   ordertotal numeric not null,
> > >   tax numeric not null,
> > >   shipping numeric not null,
> > >   subtotal numeric not null
> > >   );
> > >
> > Create the table just like above and use nextval, the sequence is only
> > autoincremented when you omit the column on insert, or insert DEFAULT.
> This pretty much clears me up,  except when you say omit the DEFAULT
> setting do you mean on the create table call, or in the insert?

no, omit the column on the insert

insert (customer_id) values (1) will auto increment the serial, as will
insert (order_id, customer_id) values (DEFAULT, 1);

> >
> > > ???
> > >
> > > 2)How can you guarantee that between the select currval('sequence')
> > > and the insertion that another value hasn't been inserted?
> > The server does this for you.
>
> How can the server know this?  What does it base it on?
when nextval is called it remembers the value.

try it. open two windows with psql do a select nextval( 'sequence' ) in
each, then do  select curval('sequence') in each


>
> Any thoughts on the date thing?
I need some time to debug, and I am working on something else at the
moment, will get to it before tomorrow.

> thanks
> ryan
>
>
>
> > >
> > >
> > >
> > >
> > > -Ryan
> > >
> > >
> > > > Dave
> > > > >
> > > > >
> > > > > thanks a ton
> > > > > ryan
> > > > >
> > > > >
> > > > > ---------------------------(end of broadcast)---------------------------
> > > > > TIP 4: Don't 'kill -9' the postmaster
> > > > --
> > > > Dave Cramer <Dave@micro-automation.net>
> > > >
> > > >
> > > > ---------------------------(end of broadcast)---------------------------
> > > > TIP 4: Don't 'kill -9' the postmaster
> > > >
> > --
> > Dave Cramer <Dave@micro-automation.net>
> >
> >
> > ---------------------------(end of broadcast)---------------------------
> > TIP 2: you can get off all lists at once with the unregister command
> >     (send "unregister YourEmailAddressHere" to majordomo@postgresql.org)
> >
--
Dave Cramer <Dave@micro-automation.net>