The only case I can think of where view partitioning makes more sense is
if it's list partitioning where you can also drop a field from your
tables. IE: if you have 10 projects, create 10 project_xx tables where
xx is the ID of the project, UNION ALL them together in a view, and
create rules on tha
On 10/13/06, Ron Johnson <[EMAIL PROTECTED]> wrote:
Hi,
I've gotten preliminary approval to buy a server and load a *lot* of
data into it. One table will eventually have 4.5Bn 330 bytes rows,
the other 9Bn 300 byte rows. Other will "only" have a billion rows.
They are easily partitioned by yy
Ron,Even though using a view means that it would have to be recreatedeach period as the oldest table is dropped,
please keep in mind: views are not really "created" ... also the command is named "create view"VIEWS, at least in PostgreSQL (and Oracle) are nothing else then "macros" for Queries - th
Hi,
I've gotten preliminary approval to buy a server and load a *lot* of
data into it. One table will eventually have 4.5Bn 330 bytes rows,
the other 9Bn 300 byte rows. Other will "only" have a billion rows.
They are easily partitioned by mm, which we call FISCAL_PERIOD.
(In fact, the app
