Tom Lane wrote:
> It won't happen. See the many, many prior discussions of this point in
> the archives.
Hmm, I searched the archives both in my local mailbox and on Google,
could you give me some pointer to a relevant thread? I certainly
want to learn from past discussions.
So, it's that MVCC issue? I will then pursue distributing indexes
and tables over more spindles and will also consider CLUSTERing the
table (although that's difficult for two *different* use cases I
have.)
However, when everything is said and done, the "it won't happen"
still sounds frightening to me. It puts PostgreSQL at a BIG disadvantage,
limiting it to 150 records per second random access ceiling when the
data needed is really all available for sequential index scan.
I'm eager to learn more about it and may be put versioning into the
index rows (or whatever it is.) Who am I to doubt Tom Lane's words
"it won't happen"? But isn't this a very serious conclusion? In any
of this we have that challenge/competition with Oracle in mind, want
it or not. If we want to use PostgreSQL seriously for our medical
record, an "it won't happen" answer can make a huge negative impression
on those I am working with. PostgreSQL for our medical record system
-- it won't happen!? ... Heck no! This can't be the answer!
regards
-Gunther
--
Gunther Schadow, M.D., Ph.D. gschadow@regenstrief.org
Medical Information Scientist Regenstrief Institute for Health Care
Adjunct Assistant Professor Indiana University School of Medicine
tel:1(317)630-7960 http://aurora.regenstrief.org