|
|
Posted by Matik on 09/27/05 19:35
Hi,
I have a small theoretical issue.
I have one table, which is prettyu large. There is lot of evaluations
running on this table, that's why, each process need to wait for
another to be finished. Sometimes, for some critical functions, it
takes to long time.
I don't think that I can speed up processes, by changing the indexes on
the tables (to increase scan time for example), because this is
something what I was experimenting with already, and it was not enought
good.
My question is, will it improve performance, if I will create second
table, exactly like this one, and I will split some evaluations, that
the one, which defenately need to run on the source table will run on
the first one, and the second evaluations, will run on the other one.
To keep data consistance between this two tables, I was thinking baout
trigger on insert on the mother table, which will transport the data to
another one.
Second part is: to improve selects on the table, should I set indexes
with option of Fill factor as possible close to 100% or as possible
close to 0%. Or maybe should I set the pad index option?
What about clustered indexes. Is it better to use them if I would like
to increase performace for selects?
Thanks in advance
Mateusz
Navigation:
[Reply to this message]
|