[
https://issues.apache.org/jira/browse/CAMEL-7904?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14218116#comment-14218116
]
Camel Guy edited comment on CAMEL-7904 at 11/19/14 5:58 PM:
------------------------------------------------------------
The main problem is that PostgreSQL doesn't let you specify the schema on the
connection properties.
Please see:
http://postgresql.nabble.com/Search-path-in-connection-string-td5718440.html
Even the beta 9.4 JDBC driver will not have this capability. Therefore, I think
using Postgres with a connection pool is just a headache, especially since my
app determines the database name, schema, etc. at run time. I use recipientList
and dynamically generate the "to" jdbc: url.
I guess for now, I'd better turn off connection pooling.
was (Author: camelguy):
The main problem is that PostgreSQL doesn't let you specify the schema on the
connection properties.
Please see:
http://postgresql.nabble.com/Search-path-in-connection-string-td5718440.html
The beta 9.4 driver will not add this feature. Therefore, I think using
Postgres with a connection pool is just a headache, especially since my app
determines the database name, schema, etc. at run time. I use recipientList and
dynamically generate the "to" jdbc: url.
I guess for now, I'd better turn off connection pooling.
> JDBC component: Setting the PostgreSQL schema search path
> ---------------------------------------------------------
>
> Key: CAMEL-7904
> URL: https://issues.apache.org/jira/browse/CAMEL-7904
> Project: Camel
> Issue Type: New Feature
> Affects Versions: 2.13.0
> Reporter: Camel Guy
>
> For PostgreSQL it is common to run two commands in a single transaction in
> order to set the schema search path:
> set search_path to foo, public;
> select 5;
> Due to connection pooling it is desirable to combine both statements in a
> single operation.
> These types of queries don't work with the JDBC component because it only
> returns the result set for the first command (set search_path), which is
> always empty.
> The easiest solution would be an option to skip the first N result sets.
> However, if any of those result sets reported an error, an exception should
> be thrown.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
