[BioSQL-l] Oracle support...
Hilmar Lapp
hlapp at gnf.org
Thu Jul 31 23:21:35 EDT 2003
On Thursday, July 31, 2003, at 09:48 PM, Len Trigg wrote:
> Hilmar Lapp wrote:
>> Is your Biojava/Oracle support for biosql able to handle names
>> different from the otherwise expected Pg schema?
>
> I just thought of another question. Does the BioPerl/Oracle support
> for BioSQL handle table and field names that are the *same* as the pg
> schema (i.e. can BioPerl be used to access the mysql-port schema that
> I made, or does it assume the details of the Symgene schema). I don't
> really use perl myself, but it is conceivable that someone in our
> project might want to later on.
>
It expects the synonym-based API.
I'd be willing to make a speed comparison with the view-based API and
switch it over speed is comparable.
Theoretically, full selects should be equal in speed. I'm not sure to
what extent there is a difference for joins and index-key selects, and
in 9i there may be no difference anymore, but I'd rather see it
(unit-)tested before I believe it.
The difference at least up to 8i was that selects from a view will
always read the row from the table (and then return only those columns
that you requested), whereas in a select from a table if the select
list can be satisfied from the indexes that were used it wouldn't even
read from the table. E.g. if you selected only the primary key from a
table, Oracle would only open the primary key index, but not the table.
Bioperl-db doesn't do partial row selects. There are some limited joins
though (which is the same situation: those tables that don't appear in
the select list can usually be satisfied from the index alone).
As I said, the difference if any may be minuscule. I'd just rather know
before switching.
And at any rate I guess it's clear that that wouldn't happen now.
-hilmar
> Cheers,
> Len.
>
--
-------------------------------------------------------------
Hilmar Lapp email: lapp at gnf.org
GNF, San Diego, Ca. 92121 phone: +1-858-812-1757
-------------------------------------------------------------
More information about the BioSQL-l
mailing list