Re: Locks rows

From: Pablo Sanchez <pablo_at_dev.null>
Date: Mon, 20 Jan 2003 15:50:26 -0600
Message-ID: <Xns930996FBEE82Cpingottpingottbah@216.166.71.233>

"Howard J. Rogers" <howardjr2000_at_yahoo.com.au> wrote in news:ZDZW9.28988$jM5.74139_at_newsfeeds.bigpond.com:

> Who can say? His post says he wants to stop two people simultaneously
> *selecting*.

Correct! I was reading into it that perhaps he was needing to implement SERIALIZABLE access. But you're right, I have no idea so was hazzarding (probably incorrect) a guess. :)

> But that's an application coding issue. If he's talking about
> selecting via SQL*Plus, then the select for update trick isn't going
> to work, no?

Not sure what you mean 'cuz in SQL*Plus, I can issue a 'select for update' in one shell and do the same command in another shell and one blocks the other -- as expected.

However, as you point out, if one of the accesses is a 'select' rather than a 'select for update' I can access the data.

On a related note, plz help me understand something that I've been meaning to figure out but since I have your ear (okay, eyes really!) I'll run it by ya .. Oracle is tracking the 'select for update' in the data block (INIT TRANS right?). Is that information actually flushed to the redo log or are INIT TRANS in-memory buckets set aside to track the select for updates?

Thx!

-- 
Pablo Sanchez, High-Performance Database Engineering
http://www.hpdbe.com

Received on Mon Jan 20 2003 - 15:50:26 CST