Re: [GENERAL] strange commit behavior

2008-11-18 Thread Flavio Palumbo
Hi Thomas, I wrote the following test case that seems to work fine. Any suggestion ? Thanks a lot Flavio import java.sql.Connection; import java.sql.PreparedStatement; import java.sql.Savepoint; public class TestCommitPostgres { public TestCommitPostgres() { java.io.BufferedRead

Re: [GENERAL] strange commit behavior

2008-11-18 Thread Grzegorz Jaƛkiewicz
On Tue, Nov 18, 2008 at 1:25 PM, Craig Ringer <[EMAIL PROTECTED]>wrote: > > > You can use savepoints (see the manual) to recover from errors, though > they have some issues if you use hundreds of thousands of savepoints in > a single transaction. It's much better to check the data on INSERT to > m

Re: [GENERAL] strange commit behavior

2008-11-18 Thread Craig Ringer
Flavio Palumbo wrote: > Unfortunately postgres seems to work in a different way, cause if there is > just one error while the transaction is active I'm not able to commit the > well formed data in the db, no matter if the good records were inserted > sooner or later the error. Yes, that's right.

Re: [GENERAL] strange commit behavior

2008-11-18 Thread Albe Laurenz
Flavio Palumbo wrote: > I developed a little tool in Java that updates databases throught text > files. > > In this tool there is an option that allows the user accepts a defined > amount of errors and save the well formed data. > > To do this I start commitment control when the process begins >

Re: [GENERAL] strange commit behavior

2008-11-18 Thread Thomas Kellerer
Flavio Palumbo, 18.11.2008 10:01: I tested this tool under MySql and Oracle and everything went as expected. Unfortunately postgres seems to work in a different way, cause if there is just one error while the transaction is active I'm not able to commit the well formed data in the db, no matter