Re: [HACKERS] Re: [INTERFACES] ODBC is slow with M$-Access Report - Mailing list pgsql-hackers

From Bruce Momjian
Subject Re: [HACKERS] Re: [INTERFACES] ODBC is slow with M$-Access Report
Date
Msg-id 199806021621.MAA12356@candle.pha.pa.us
Whole thread Raw
In response to Re: [INTERFACES] ODBC is slow with M$-Access Report  ("Jose' Soares Da Silva" <sferac@bo.nettuno.it>)
Responses Re: [HACKERS] Re: [INTERFACES] ODBC is slow with M$-Access Report
List pgsql-hackers
>
> We are working on a project that IMHO give more prestige to
> PostgreSQL.
> The Hygea project concern the use of an Unix-like Operating  sys-
> tem  as  "back-end" of a Client M$-windows application connected
> by ODBC that will be installed in about 80 Italian Helth Depart-
> ments for the veterinary controls and prevention.
> Therefore...
>
> O.S.: We choose Linux for his proved reliability.
>
> Client: We choose to develop the Client with M$-Access because we
> need (unfortunately) a complete integration with Micro$oft World.
>
> Database: We choose PostgreSQL for his reliability  and  for  his
> compatibility with SQL/92 standard recommendation and for his ex-
> cellent technical support provided by "The PostgreSQL Development
> Team" and his mailing lists.

Great.

>
> Nevertheless  the  union  among M$-Access and PostgreSQL is quite
> suffered for the following reasons:
>
> 1. The PostgreSQL doesn't use the index with  "OR"  operator  and
> so is not possible to define a multiple key to use with M$-Access
> and we need to retreat using OID as primary keys (thanks to Byron
> Nikolaidis and David Hartwig of insightdist.com that are doing a
> really great job with ODBC driver), but with the obvious consequences.

Yes, we need to work on this.  I am sure performance really suffers
because of this.  Vadim, is this on your short list?

>
> 2.  As PostgreSQL doesn't allow an "ORDER BY" on columns not
> included in the target list of the "SELECT", (I know that it is
> SQL/92 standard, but IMO it's a fool thing), therefore, is not possible
> to  have the "dynaset "sorted for any field that is different from
> the key (in our case the useless OIDs).

David at Insight just added this, so it certainly will be in 6.4.

>
> 3. The times required to run complex reports (for example those that
> include LEFT JOINS) is very long (about 15 minutes to retrieve
> 2850 rows).

Yea, we need this too.  Not sure where we are with this.  Can you give
an example?



--
Bruce Momjian                          |  830 Blythe Avenue
maillist@candle.pha.pa.us              |  Drexel Hill, Pennsylvania 19026
  +  If your life is a hard drive,     |  (610) 353-9879(w)
  +  Christ can be your backup.        |  (610) 853-3000(h)

pgsql-hackers by date:

Previous
From: David Hartwig
Date:
Subject: Re: [INTERFACES] ODBC is slow with M$-Access Report
Next
From: The Hermit Hacker
Date:
Subject: Re: [HACKERS] Re: [INTERFACES] ODBC is slow with M$-Access Report