Re: [GENERAL] Prepared statement performance...

Thank you for your remark. But we're talking about SQL statements (and not
JDBC statements), aren't we? I have not found anything as yet which would
indicate that the JDBC API significantly redefined the meaning of the term
"statement" compared to what had been known as SQL statement well before
JDBC appeared. And in this regard (i.e. to clarify what has/had been meant
by "SQL statement"), the SQL spec seemed to me a good starting point,
although anecdotical evidence can, of course, also be used, if you can
provide some.

Your remark is correct, if you mean that the JDBC API does not explicitely
restricts you from putting in your SQL statement anything you wish. But
(provided I understand the big picture) JDBC has been designed for a very
specific kind of resource managers: database servers communicating with
their clients via SQL statements. There exists a specification for a very
generic JAVA resource adapter, which explicitely allows you to put over the
wire anything you wish (actually, anything the resource manager it is
intended for understands). But this generic resource adapter is called
Connector, and not JDBC driver. So if you want to interpret liberally the
API between the server and the client, the postgres java client classes (in
what is now called JDBC driver) should implement the "connector" interfaces
rather than the "sql" interfaces.

And as to the whereabouts of the JDBC API spec, a good place to look for it
is www.java.sun.com .

Cheers,

Peter


----- Original Message -----
From: "Tom Lane" <tgl@sss.pgh.pa.us>
To: "Peter Kovacs" <peter.kovacs@sysdata.siemens.hu>
Cc: "Barry Lind" <barry@xythos.com>; "Dmitry Tkach"
<dmitry@openratings.com>; <pgsql-general@postgresql.org>;
<pgsql-jdbc@postgresql.org>
Sent: Saturday, September 28, 2002 6:25 PM
Subject: Re: [JDBC] [GENERAL] Prepared statement performance...


> "Peter Kovacs" <peter.kovacs@sysdata.siemens.hu> writes:
> > ... but the SQL 92 spec should be, of course, the
> > ultimate authority in this regard.
>
> The SQL spec has never heard of JDBC.  I assume there is a spec
> somewhere for the JDBC API, but it's not SQL92 nor SQL99.
>
> regards, tom lane
>