[ZODB-Dev] ZODB and ORBit-Python problems

[Christian Robottom Reis]

On Mon, 21 May 2001, Michel Pelletier wrote:

> But not all of them at once.  You can throw away the employee object after
> you are done sucking the statistical life-blood out of it.

Quite true. While this does help memory bloat, however, it does not avoid
having to touch all objects of that class to obtain the data. Which is,
AFAICS, unavoidable, unless we autocalculate a field as sales are
generated.

In this area would a query language and structured storage make ZODB
excel, but from the lack of response it seems like a whole separate
category of TODO items -- Nice-To-Have-But-Kinda-Remote-Possibility :-)

(Jim, does this have to do with reliance on cPickle?)

> I'm not sure if there is any clear performance win with not loading an
> object into memory after reading it, either way you do the IO, and that's
> the slow part.

Well I think (but am not sure) it depends on how the data is organized on
disk; relational databases can get off very quickly because they don't do
any data conversion and they store data in a structured form, thus making
reads (and summaries of the data read) very fast. So apart from doing less
IO, less time is spent doing data conversion and presentation. At least
AFAICS.

Take care,
--
/\/\ Christian Reis, Senior Engineer, Async Open Source, Brazil
~\/~ http://async.com.br/~kiko/ | [+55 16] 274 4311