On Tuesday 08 October 2002 09:54 pm, Gerard van de Looi wrote:
> Hi,
>
> I produced a small document on the advantaged and disadvantages of using
> prepared statements and a prepared statement cache to improve the prepared
> statement performance.
> I'm working on an implementation of a prepared statement cache that affects
> the current multiconnection pool only at a few methods.
sounds good.
Any results yet?
As far as I understand in MMBase you almoost always need 2 query get a node
(one to find out in what database table the node is stored and on to actualy
find the node). It thing this migth be a problem, is it possible to define
the table to query in the PreparedStatement after the statement was created?
do you indent to create prepared statment for stuff like that.or more on a
higher level?
Are you targeting a special database?
There are 3 more things that might be of interest to you
1- The security project that would very much like to enhance performance by
delegating security rules to the database (select bigquery where
bigcontratins and owner in "users")
2-(this was discussed on irc but not on the list yet) an other big speed
improvement could be gained by creating "koppel tabellen" for every possible
relation. so creating a table for relations between "user and images of type
posrel). by doing this nodes could be retrived in one query.
3-The Query project is working on an abstract way to define the "path" to
follow when issuing a query (typicaly multilevel
person->posrel->images->related->urls
->related->documents
)
It quite possible that it might help to define the "prepared" statement.
anyway.. whaterver it's that your doing.. and whatever your goals are..
it's just me "talking" so you don't need to listen :).
success!
