Comments on: Serializable transactions are NOT single threading a database

By: murat yücel

murat yücel — Mon, 08 Dec 2008 09:07:37 +0000

thanks a lot it helped me so much

By: Bernd

Bernd — Wed, 30 Apr 2008 14:10:55 +0000

Good article - but I still can’t solve my problem Maybe you can help me with that.

I have an XML Document stored in an Oracle DB. There are 4 CF functions (delete,rename,insert, update) to make changes in the XML document. I need to make sure that while one function changes the document no other function can alter the document. As there are a couple of CF Servers I can’t use CF lock. The lock needs to take place in the DB. How can I do that?

Thanks in advance

Bernd

By: Jochem

Jochem — Sat, 29 Mar 2008 18:35:24 +0000

If you are looking for the best book on a specific database, start with the manual. Just read it cover to cover.
If you are looking for something that is not specific to a particular database, I would recommend starting with Architecture of a Database System. If you want to get really theoretical you could try Concurrency Control and Recovery in Database Systems.

But what you should never forget is that you can just try things. You can just start a query editor and write queries. And then start another one and see how they interact.

By: T1

T1 — Sat, 29 Mar 2008 13:44:42 +0000

What is the best research book on this subject?

By: Tom Kitta

Tom Kitta — Thu, 28 Feb 2008 00:04:29 +0000

Thanks a lot for informative post. I just discovered a problem with one of our DBs where I have supposedly rare phantoms show up once per about 10000 transactions. Query A checks whatever user response was recorded in DB, if it was not query B inserts it. Once in 10000 times query A says “nothing in DB” while an insert is pending. B initiates another insert and I end up with two records that have identical data and identical time stamp. I am hoping placing A and B in serialize block will solve the issue.

By: Brad Wood

Brad Wood — Tue, 12 Feb 2008 16:18:52 +0000

Thanks for the good explanation of transaction isolation levels. This is only something I have fairly recently come to grasp. Good, understandable commentary isn’t always around to help.