From: | "Peter Kovacs" <maxottovonstirlitz(at)gmail(dot)com> |
---|---|
To: | "Heikki Linnakangas" <heikki(at)enterprisedb(dot)com> |
Cc: | pgsql-jdbc(at)postgresql(dot)org |
Subject: | Re: PreparedStatement rounds doubles to scale 14 during update |
Date: | 2007-09-05 11:42:00 |
Message-ID: | b6e8f2e80709050442s7439d8ebt345024599931400d@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-jdbc |
Heikki,
Thank you for your reply.
What are base-10 datatypes? Are they numeric types? I cannot find them
in the backend documentation.
Is there a mapping of Java types to backend types documented somewhere?
Thanks a lot
Peter
On 9/5/07, Heikki Linnakangas <heikki(at)enterprisedb(dot)com> wrote:
> Peter Kovacs wrote:
> > Hi,
> >
> > I have the following problem with postgresql-8.1-407.jdbc3.jar used
> > against a 8.1 version backend.
> >
> > A table "chemterms_cols_test" has a column "logp" defined as:
> >
> > logp numeric(30,17)
> >
> > The double value 6.118992224252588 gets rounded to 6.11899222425259000
> > during updates with PreparedStatements using place-holder. (There is
> > no rounding with Statement and the value specified in the SQL
> > literally.) Is this the expected behaviour?
>
> The problem is in the conversion from double to numeric in the server.
> The server-side float8 data type doesn't have enough precision for that
> many decimal places. Because converting between numeric and
> floating-point data types almost always leads to rounding errors at some
> point, I would strongly recommend sticking to base-10 data types
> everywhere if you care about the precision.
>
> If you use BigDecimal instead of double in your java-program, it works fine.
>
> --
> Heikki Linnakangas
> EnterpriseDB http://www.enterprisedb.com
>
From | Date | Subject | |
---|---|---|---|
Next Message | Heikki Linnakangas | 2007-09-05 12:07:59 | Re: PreparedStatement rounds doubles to scale 14 during update |
Previous Message | Albe Laurenz | 2007-09-05 10:34:14 | RE: [JDBC] Questión of JDBC |