Обсуждение: hi problem with installing postgresql8.1

Hi all,
    I am new to pgsql family.I had a problem with installing postgres8.1 i was getting the following error at the 4th step of installation the error is as follows:

Data directory error:
  The specified data directory is not empty

   If you have an existing database with teh same major version number,you do not need to initialise a database cluster.
   If you have an existing database with the same major version you need to backup your old database and create a new one.

   why this warning actually arises.If  I install by unchecking the initialise database cluster then the next three steps of installation procedure is not shown.It goes directly installation and starts installation after some time a fatal error arises saying accont not matching....................
    and rolls back.i have tried it with lots of account names and lot of super user names.The steps i was talking about are as per the given link.

              http://pginstaller.projects.postgresql.org


    Can any one give the solution to this problem.Please suggest solution to overcome this problem ASAP.

Thanks in advance,
Venu.

Based on a couple of other posts, I think I'm out of luck, but I'm hoping
something might have changed recently.

I'm loading a very high volume of data with COPY using libpq - about 100+
million rows per day. The problem is that the COPY sometimes aborts with

  invalid UTF-8 byte sequence detected

I'd prefer not to add to the overhead by pre-validating every string, since
PostgreSQL validates it already.  Is there a way to get the server to
replace invalid characters with something like blank instead of generating a
fatal error?

Wes

If you don't care to store those characters then maybe you should change
the database character set to use SQL_ASCII instead of UTF-8.  I believe
ASCII will quietly discard those characters when converting from UTF-8.

Mike

 On Wed, 2006-03-15 at 23:20 -0600, Wes wrote:
> Based on a couple of other posts, I think I'm out of luck, but I'm hoping
> something might have changed recently.
>
> I'm loading a very high volume of data with COPY using libpq - about 100+
> million rows per day. The problem is that the COPY sometimes aborts with
>
>   invalid UTF-8 byte sequence detected
>
> I'd prefer not to add to the overhead by pre-validating every string, since
> PostgreSQL validates it already.  Is there a way to get the server to
> replace invalid characters with something like blank instead of generating a
> fatal error?
>
> Wes
>
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 1: if posting/reading through Usenet, please send an appropriate
>        subscribe-nomail command to majordomo@postgresql.org so that your
>        message can get through to the mailing list cleanly

On 3/15/06 11:42 PM, "mike" <mike@thegodshalls.com> wrote:

> If you don't care to store those characters then maybe you should change
> the database character set to use SQL_ASCII instead of UTF-8.  I believe
> ASCII will quietly discard those characters when converting from UTF-8.

I thought about that, but I do want to allow UTF-8 to be stored.  I just
want it to replace illegal characters with some valid character so that
invalid records will load.   Even if I preprocess the data, I can do no more
than that.

Wes

On Mar 16, 2006, at 14:42 , mike wrote:

> If you don't care to store those characters then maybe you should
> change
> the database character set to use SQL_ASCII instead of UTF-8.  I
> believe
> ASCII will quietly discard those characters when converting from
> UTF-8.

SQL_ASCII is *not* ASCII. See the "What's a good default encoding?"
thread on this same list from today.

http://archives.postgresql.org/pgsql-general/2006-03/msg00685.php

I don't believe it will discard anything on import if the database is
SQL_ASCII encoded.

Michael Glaesemann
grzm myrealbox com

On 3/16/06 12:13 AM, "Michael Glaesemann" <grzm@myrealbox.com> wrote:

> SQL_ASCII is *not* ASCII. See the "What's a good default encoding?"
> thread on this same list from today.
>
> http://archives.postgresql.org/pgsql-general/2006-03/msg00685.php
>
> I don't believe it will discard anything on import if the database is
> SQL_ASCII encoded.

That might be worth a shot.  I don't really understand the ramifications,
though, especially given Tom's warning.  I guess as long as I don't care
about things like sort order for those fields, it may not matter much.

Wes

On Mar 16, 2006, at 15:39 , Wes wrote:

> That might be worth a shot.  I don't really understand the
> ramifications,
> though, especially given Tom's warning.  I guess as long as I don't
> care
> about things like sort order for those fields, it may not matter much.

I hope you didn't take my comments as a suggestion: I think you
*should* preprocess your data and use UTF8 as the default encoding
(or another encoding that includes all of the characters you hope to
use) for your database.

Michael Glaesemann
grzm myrealbox com

On 3/16/06 12:41 AM, "Michael Glaesemann" <grzm@myrealbox.com> wrote:

> I hope you didn't take my comments as a suggestion: I think you
> *should* preprocess your data and use UTF8 as the default encoding
> (or another encoding that includes all of the characters you hope to
> use) for your database.

Yeah, after thinking about it some more, it probably isn't such a good idea
- it would make the database act rather strange with non U.S. Data.  I
really hate to have to burn the cycles to validate anything that looks like
extended UTF-8 characters, only to have Postgres do it again.  But maybe
there's no choice.

Wes