Re: Unicode database + JDBC driver performance

On Sat, Dec 21, 2002 at 11:45:08AM -0500, Tom Lane wrote:
> Jan Ploski <jpljpl@gmx.de> writes:
> > The Unicode database is slower both on INSERTs and especially on
> > SELECTs.
>
> What did you have PGCLIENTENCODING set to in each case?  I am wondering
> if you were incurring an extra conversion ...

Tom,

When connecting to the Unicode database, I did not use any 'charSet'
parameter for the JDBC driver (which I assume translates to
PGCLIENTENCODING on the backend?). When I specified charSet=UNICODE,
I got an exception when executing the query. Without, it worked fine.

For the LATIN1 database, I used charSet=LATIN1 in driver's connection
string. It worked as expected (the conversion of non-ASCII characters
was not correct, but that's ok).

Is there some other way to find out what PGCLIENTENCODING is used
in each case?

Take care -
JPL