| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Mailing Lists -> Oracle-L -> Re: Index-Organized Table experiences
There doesn't seem to be any real point
in using IOTs for your scenario. For an
intersection table that small, the saving in
space is likely to be non-existent. You may
as well have:
create table intersection(
proj_id typeX,
emp_id typeX,
constraint int_pk primary key (proj_id, emp_id)
);
Given the space overheads in indexing and IOTs in particular, (even allowing for the documented reduced size of secondary indexes in 10g), I doubt if the table plus two indexes will take up much more space than the IOT plus secondary. (In fact, in my test cast with, the t + 2i used a little less space than the IOT + s).
Since secondaries have had a revamp in 10g, I'd be cautious about assuming that they would do all the right things.
Since you only need to access an index to traverse between the employee and project tables, irrespective of direction, the table would only get into memory when you were inserting or deleting rows.
Stick with the traditional technology when the new technology has no apparent added value.
One minor detail with IOTs - if two users try to update the same row, the second user's TX wait will be mode 4, rather than mode 6.
Regards
Jonathan Lewis
http://www.jlcomp.demon.co.uk
The Co-operative Oracle Users' FAQ
http://www.jlcomp.demon.co.uk/faq/ind_faq.html
April 2004 Iceland http://www.index.is/oracleday.php June 2004 UK - Optimising Oracle Seminar July 2004 USA West Coast, Optimising Oracle Seminar August 2004 Charlotte NC, Optimising Oracle Seminar September 2004 USA East Coast, Optimising Oracle Seminar September2004 UK - Optimising Oracle Seminar
We are looking to implement IOTs for a couple of intersection entities in a 10g db. I would like to hear from those brave enough to actually use IOTs what is the good, the bad and the ugly.
example:
Employee (heap table)
Project (heap table)
There is a many-to-many relationship between the tables (1 employee can be on many projects and 1 project can have many employees).
The emp_project table is the intersection entity containing emp_id and project_id as the only columns. There are FK constraints on each of the columns. The combination of emp_id and project_id is unique.
This situation *sounds* like the right one for an IOT, otherwise we would have 1 table and 2 indexes (1 on each column).
My main concerns are:
1) Integrity/performance
2) Locking behavior (do I need to adhere to the traditional
"index all foreign keys" rule to prevent excessive locking?)
3) Any especially nasty gotchas
Thanks,
Daniel
 |
 |