Re: Best way to make query/update sequence atomic

From: Thomas Kyte <tkyte_at_us.oracle.com>
Date: Fri, 27 Aug 1999 11:21:02 GMT
Message-ID: <37cc7447.10690191@newshost.us.oracle.com>

A copy of this was sent to "Adrian Bowen" <adrian_at_raptor.win-uk.net> (if that email address didn't require changing) On Fri, 27 Aug 1999 08:57:26 +0100, you wrote:

>I have a PL/SQL stored procedure which does a SELECT to test for the
>existence of a row satisfying a set of criteria, then either updates the
>existing row or inserts a new one. How do I ensure that two concurrent
>sessions don't both get 'no row found' and try to do the insert? Is there a
>standard technique for doing this? At present the only solution I can see is
>to use DBMS_LOCK to ensure that only one instance of the SP is executing at
>any time.
>
>Apologies if this is an entry-level question - any suggestions appreciated.
>
>Adrian Bowen
>

why not something like:

  ...
  begin

     insert into T values ( ..... );
  exception

     when dup_val_on_index then
        update t set ......;

  end;
  ....

or

  update T set .....
  if ( sql%rowcount = 0 )
  then

     insert into t values .....
  end if;

don't do the select, try inserting it -- if it fails row already exists so update it

or

try updating it, if you don't find it insert it.

--
See http://govt.us.oracle.com/~tkyte/ for my columns 'Digging-in to Oracle8i'... Current article is "Part I of V, Autonomous Transactions" updated June 21'st  

Thomas Kyte                   tkyte_at_us.oracle.com
Oracle Service Industries     Reston, VA   USA

Opinions are mine and do not necessarily reflect those of Oracle Corporation Received on Fri Aug 27 1999 - 06:21:02 CDT