[sqlalchemy] Re: unexpected exception - Manlio Perillo - com.googlegroups.sqlalchemy

7 messages in com.googlegroups.sqlalchemy[sqlalchemy] Re: unexpected exception

From	Sent On	Attachments
Manlio Perillo	21 Feb 2007 13:00
Michael Bayer	21 Feb 2007 15:27
Manlio Perillo	22 Feb 2007 03:03
Michael Bayer	22 Feb 2007 10:04
Manlio Perillo	22 Feb 2007 10:14
Michael Bayer	22 Feb 2007 12:18
Manlio Perillo	22 Feb 2007 12:24

Subject:	[sqlalchemy] Re: unexpected exception
From:	Manlio Perillo (manl...@gmail.com)
Date:	02/22/2007 12:24:32 PM
List:	com.googlegroups.sqlalchemy

Michael Bayer ha scritto:

On Feb 22, 2007, at 1:14 PM, Manlio Perillo wrote:

Michael Bayer ha scritto:

On Feb 22, 6:04 am, Manlio Perillo <manl...@gmail.com> wrote:

Michael Bayer ha scritto:

its what it says...its updating the database and not finding the rows it expects, indicating some other concurrent process already deleted that row.

Is this actually possible, even if each query run in its own transaction? The database is Postgresql 8.1, Debian Etch.

sure...its a deleted row so row locking doesnt even count....both transactions can occur totally separate from each other too.

Ok, but then why SQLAlchemy raises an Exception?

because the second transaction tried to do something, like delete a row, and the row wasnt there....some other transaction already got to it. therefore the operation failed. this is just basic "optimistic concurrency checking".

The problem is the operation is not failed. All the rows that must be delete have been deleted.

Suppose I'm going to delete 10 rows, and got a count of 8 rows. If SQLAlchemy raises an exception, the transaction will be rolled back and 8 rows will not be deleted at all.

I'm really not sure about the right thing to do here.