ibatis-user-java mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Philippe Laflamme <philippe.lafla...@mail.mcgill.ca>
Subject Re: Oracle + iBATIS: more details
Date Wed, 16 Mar 2005 17:38:06 GMT

I've experienced differences between JDBC executed vs Toad executed 
statements. If I remember correctly, our issue at the time was due to 
differences in connection variables: Toad and the JDBC connection did 
not have exactly the same connection settings. You might want to look at 
connection variables to make sure they're identical.

You probably know all of this but here are a few suggestions...

What you first need to do is isolate the problem *outside* of your web 
app. There are too many differences between a WebApp executed statement 
vs Toad...

I'd setup the statement in these environment:
1- Toad
2- SQL*Plus
2- Plain JDBC
3- iBatis only (no webapp)

Using SQL*Plus will allow you to know if Toad is doing something fancy. 
Your Plain JDBC setup will allow you to know if your WebApp is altering 
your JDBC connections or statements, etc.

You'll waste a great amount of time if you keep comparing high level 
executed query (web, java, iBatis, jdbc) vs a very low level query (Toad 
uses native Oracle client librairies): there are just to many 
differences between the two...

Also, you state that the statement "will not complete until the first 
one does". Is the first statement a SELECT statement or are you altering 
the data (UPDATE, INSERT, etc.)? If so, play with concurency settings on 
your connection (java.sql.Connection.TRANSACTION_*)...

Hope this helps,

Brian Barnett wrote:
> We execute a query through the web app that takes 55 seconds to complete.
> While this query is executing, we execute another query against the same
> table in a different browser. It normally takes a few seconds, but it will
> not complete until the first one does.
> If we do the same thing in two instances of TOAD (sql query tool), the second
> query returns immediately.
> Can someone shed some light on what is going on, where to look, etc.?
> Thanks,
> Brian Barnett

View raw message