RE: Joint Prefetches

From: Andrus Adamchik (andru..bjectstyle.org)
Date: Mon Mar 07 2005 - 09:53:04 EST

  • Next message: jira-norepl..bjectstyle.org: "[OS-JIRA] Created: (CAY-286) Generated classes do not have package statement"

    It depends on how you look at it. I am sure we can be smarter about all
    these things. I need to analyze this a bit deeper. E.g. problem #1 (result
    set excluding certain root objects) is the same issue that happens when
    you have a join specified in qualifier, so once we implement outer join
    semantics, this will address those two seemingly unrelated issues..

    What bothers me is inconsistency - I want a default solution that works
    the same way for both multi-query and single query prefetches... or at
    least an easy way to explain these differences without going to deep into
    relational behavior...

    And by the way, I think that work done on the joint prefetches will help a
    lot with regular flattened prefetches implementation - something I am
    planning to look at shortly.

    Andrus

    > Are the limitations all temporary or permanent due to the nature of it?
    >
    > This sounds super cool. Will start playing around with it!
    >
    > Cris
    >
    >> -----Original Message-----
    >> From: Andrus Adamchik [mailto:andru..bjectstyle.org]
    >> Sent: Monday, March 07, 2005 12:48 AM
    >> To: cayenne-deve..bjectstyle.org
    >> Subject: Re: Joint Prefetches
    >>
    >>
    >> On Mar 7, 2005, at 12:44 AM, Andrus Adamchik wrote:
    >>
    >> > * Qualifying root entity on an attribute of a related
    >> entity that is
    >> > joined via to-many will result.
    >> > This problem is much worse as it breaks the object graph...
    >>
    >> Oops, a piece of content was cut from the original message. It should
    >> read:
    >>
    >> * Qualifying root entity on an attribute of a related entity
    >> that is joined via to-many will result in to-many
    >> relationship list containing fewer objects than it should.
    >> This problem is much worse as it breaks the object graph...
    >>
    >>



    This archive was generated by hypermail 2.0.0 : Mon Mar 07 2005 - 09:53:06 EST