[odb-users] Using prefetch with Oracle

[Marco Craveiro]

Hi Boris,

<prefetch hack>
> Hm, I am not sure the bulk approach (with a compiler-time pragma) is
> right in this case. There we don't really have a choice since we need
> to know the "batch buffer" size.
>
> But here it is all runtime. Plus, you may want to have different
> prefetch for different queries of the same object. In fact, you
> can already customize it for queries (but not for object loads)
> by using prepared queries (Section 4.5 in the manual):
>
> 1. Create prepared query.
>
> 2. Get its statement (statement()).
>
> 3. Cast it to odb::oracle::select_statement.
>
> 4. Call handle() on the result to get OCIStmt*.
>
> 5. Set custom OCI_ATTR_PREFETCH_ROWS.
>
> 6. Execute the query.
>
> The problems with this approach are: (1) it is tedious and (2) it
> doesn't work for non-query SELECT's (e.g., database::load()). So
> perhaps the way to do it is:
>
> 1. Provide prefetch() functions on oracle::database() and
>    oracle::connection() that can be used to modify database-wide
>    and connection-wide prefetch values. Also set it to some
>    reasonable default (say 512?)
>
> 2. Provide oracle::select_statement::prefetch() to make the
>    prepared query approach less tedious.
>
> What do you think?

Yep, makes perfect sense. Let me have a go at coding this then.

Cheers
-- 
Marco Craveiro

MD, Domain Driven Consulting
about: http://about.me/marcocraveiro
blog: http://mcraveiro.blogspot.co.uk
twitter: https://twitter.com/MarcoCraveiro

That the Ideas are themselves manifestations (of the Idea-Idea) and
that the Idea-Idea is a-kind-of Manifestation-Idea—which is a-kind-of
itself, so that the system is completely self-describing— would have
been appreciated by Plato as an extremely practical joke [Plato]. --
Alan Key