[fluent-nhib] Re: Uni-directional many-to-many

2009-09-15 Thread Hudson Akridge
Furthermore, if you've got an intelligent 2nd level Caching setup, that'll just add to the performance in terms of what NH has to load across the wire from the DB, and what it can load from memory. On Tue, Sep 15, 2009 at 3:07 PM, Hudson Akridge wrote: > Lazy load :) Do a typical MTM setup with b

[fluent-nhib] Re: Uni-directional many-to-many

2009-09-15 Thread Hudson Akridge
Lazy load :) Do a typical MTM setup with both sides MTM, one marked as inverse, and set both sides to lazy load. That will proxy the collections until you use them (I am assuming that a single image is not associated to *every* document in your database). On Tue, Sep 15, 2009 at 2:56 PM, DannyT w

[fluent-nhib] Re: Uni-directional many-to-many

2009-09-15 Thread DannyT
Thanks Hudson, will revisit typical MTM and try to recall what the hurdle we were trying to overcome with that was (been going round in circles for hours so I forget!). Cheers for the thoughts Dan 2009/9/15 Hudson Akridge > Furthermore, if you've got an intelligent 2nd level Caching setup, that

[fluent-nhib] Re: Uni-directional many-to-many

2009-09-15 Thread DannyT
I guess I've manually mapped this before and fallen in the 70% bracket without issues. Is it then possible for Fluent AutoPersistanceModel to create such a mapping somehow even if it does result in 70/30 reliable code? Or what are the alternatives? The problem is this is a remote service that rec

[fluent-nhib] Re: Uni-directional many-to-many

2009-09-15 Thread Hudson Akridge
I've attempted to do this before, with extremely limited success. You question isn't a FNH one, but an NH question, and the quick answer is, no, it is not possible to do a uni-directional MTM. You can indeed map it as such, and it'll work for about 70% of the things you want to do, but you will get