Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Retrieving records in order they were placed.

Re: Retrieving records in order they were placed.

From: Mark G <someone_at_hot>
Date: Fri, 9 Jul 1999 14:18:32 +0100
Message-ID: <3785f461.0@145.227.194.253>


Why don't you build on some of your ideas.

Define 'very very high' number of transactions. A sequence can get to a very high value before it raps (sorry Wraps, no Will Smith here!), set it to something ridiculousy high.

Instead of bouncing the machine, why don't you have a job (DBMS_JOB) to drop the sequence and rebuild, starting with 1. The job could be set to run whenever you want it.

Also timestamps. I know you are inserting records in quick succession but you can try detecting timestamps down to 1/1000th of a second.

Or combine the two, timestamps and sequences. That way, if the sequence wraps, the timespamp will be well ahead of itself.

Hope one of these ideas inspires you!

M

b_rich2_at_my-deja.com wrote in message <7m4rcj$9gd$1_at_nnrp1.deja.com>...
>I have to put records into a table, and later retrieve them in the order
>in which they were placed. I have a very very large number of entries to
>be placed in a live system. The records are logs of important events in
>the system.
>
>My first thought was to use the unix time stamp to place as the field to
>order on, but there will probably be many entries entered in rapid
>succession. Thus many entries could be entered at the same time stamp
>interval.
>
>I then though of having a sequence number associated with each record,
>increasing one by one with each record stored. This will work well until
>the sequence numbers begin to wrap. If they are retrieved in order of
>sequence number, and then they wrap, the records entered after the wrap
>(sequences 1,2,3,4 etc) will all of a sudden be pulled out before the
>earlier records with high sequence numbers.
>
>Another idea is to just use the sequence numbers, and re-boot the box
>every few weeks when the sequence numbers get high, but this is really
>unacceptable in the environment where this will be run.
>
>Does anyone know how to get around this problem? Any help would be
>GREATLY appreciated.
>
>
>Sent via Deja.com http://www.deja.com/
>Share what you know. Learn what you don't.
Received on Fri Jul 09 1999 - 08:18:32 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US