Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Usenet -> c.d.o.misc -> Re: An 1-to-1 relation: Is it normal?
I have done this in some cases with good results. I look primarily at how the data will be accessed. If I have a set of data for an entity where some will be accessed frequently and some seldom, and if the seldom accessed data is not going to available anyway for a large number of the entity, I start to lean toward splitting it into 2 tables with a 1:0or1 relationship.
Even though null fields don't take up much space, there is some overhead for maintaining info about the field in the header data. Why keep it if it's not going to be used much? Splitting them can also cut down on later expansion of the row and possible chaining when the data no longer fits into the original block. More rows can be packed into a block for more efficient access The seldom used data can be put into a different tablespace if it makes sense to do so.
My point is, under certain circumstances, splitting the table can be useful. You have to analyze not only the quantity of data involved but also how it is going to be accessed.
For what it's worth
Jim Gregory
Any expressed opinions are mine and do not reflect those of my clients or employer
jeffd <jeffd_at_allensysgroup.com> wrote in article <352134DA.3493_at_allensysgroup.com>...
> Igor Sereda wrote:
>> > So what we are getting is a one-to-one-or-zero relation.
![]() |
![]() |