[Bioc-devel] Update on SQLite-based annotation data package (prototype available)

[lgautier at altern.org]

> Vincent Carey 525-2265 <stvjc at channing.harvard.edu> writes:
>
>>>
>>>   2. More flexible queries.  With the SQLite-based packages, many
>>>      queries that currently require loops over possible many entire
>>>      environments can be accomplished in one statement.  Using some
>>>      simple SQL statements, I've been able to improve the performance
>>>      of the hyperGTest function by 10x.  Focused queries will
>>>      generally be much faster with the SQLite-based packages.
>>
>> do we need a sql tutorial doc (i know there are plenty on the
>> web but perhaps some that are focused on the types of queries to
>> be used here?)  helper code that 'translates' R-like actions to
>> SQL may be feasible for some of the more common tasks.
>
> I'm hoping that an alternative API will solidify Real Soon Now.  I
> would much prefer promoting a well documented API than raw SQL.  Using
> raw SQL is effective, but relies on the schema definition.
>
> But perhaps my comments are orthogonal to your suggestion.  A SQL
> tutorial with "translations" of R concepts is a great idea.

I would only promote what Seth would prefer promoting.
Having an API could be a better solution, as it would allow to provide a
unified front-end to annotation packages while letting annotation to be
stored in a number of different backends (loaded environments like it was
the case, the coming SQLlite ones, remote SQL database, web-service,
etc...).
Having an API would also permit to make changes to the SQL schema without
causing a lot of trouble to all users.


Laurent