Home » SQL & PL/SQL » SQL & PL/SQL » initial extent - large table
initial extent - large table [message #19332] Tue, 12 March 2002 05:11 Go to next message
Christine
Messages: 18
Registered: March 2002
Junior Member
Hello ,

I need to improvre the performence of my DB.
Every 15 seconds I need to write 650 rows into a table.
(an id number(15) and a value float and some other number(1)).
When the table is empty the insert is done in less than
15 seconds but as soon as the database become bigger performence reduce.
If I am changing initial extent, next extent will I have better performence.
How do i compute then ?
thank you for the answer
Christine
Re: initial extent - large table [message #19338 is a reply to message #19332] Tue, 12 March 2002 07:48 Go to previous message
Grant
Messages: 578
Registered: January 2002
Senior Member
Changing initial extent will have no effect once the table has been created. But you are on the right track. You might recreate the table with a large initial extent. Make a best estimate on how large the table will ever get and make that your initial extent. Make your next extent is large enough so many extents will not occurr as you add more data.

Extents no longer have a big impact on performance but it still has to do work to extend.

What is more likely the problem is the number of indexes you have on the table. Everytime you insert a record all indexes are updated. Take a look at that. Another issue could be locking but only if you are updating records at the same time.

This applies if your inserts are slowing down. If when you add many records and selects are slowing down you may need to add indexes and analyze statistics on your tables often so the optimizer can make better choices. If you have indexes and and your selects are slow then analyze statistics and/or use hints in your queries to use indexes. You still need to analyze stats often with that many inserts.
Previous Topic: PL/SQL
Next Topic: how to retrieve the column names from tables using PL/SQL
Goto Forum:
  


Current Time: Thu May 09 05:07:07 CDT 2024