Message-ID: <ea2a08fe7a53481ae973dadaf174dc2bf6e89a8e.camel@cybertec.at>
Subject: Re: Performance killed with FDW when using CAST.
From: Laurenz Albe <laurenz.albe@cybertec.at>
To: Jorge Torralba <jorge.torralba@gmail.com>, pgsql-admin@postgresql.org
Date: Wed, 17 Apr 2019 07:03:50 +0200
In-Reply-To: 
 <CACut7uSz_3JiJ4w80VrQYECtOAt1Ya5E=EdLBAJ2_DnYwqsjGA@mail.gmail.com>
References: 
	<CACut7uSz_3JiJ4w80VrQYECtOAt1Ya5E=EdLBAJ2_DnYwqsjGA@mail.gmail.com>
Organization: Cybertec
Content-Type: text/plain; charset="UTF-8"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Precedence: bulk

Jorge Torralba wrote:
> We setup a FDW server for a large table that has an hstore data type as on eof the coums.
> 
> When we run the query directly on the FDW server
> 
> select * from recent_events where account_id = 1 AND (attributes -> 'account_incident_id')::integer = 2617116 limit 10;
> 
> The response is 2ms . which is awesome.
> 
> When we run it from the remote server where we have the server defined, the query never returns.
> 
> Our server definition is as follows.
> 
> (host 'fdwserver', dbname 'mydb', port '5432', fetch_size '1000', use_remote_estimate 'true', extensions 'hstore')

After debugging into this, it seems that the hstore operator -> cannot
be pushed down because of collation problems.

Yours,
Laurenz Albe
-- 
Cybertec | https://www.cybertec-postgresql.com