kind1 is really class Kind1(search.SearchableModel).  The query
against Kind1 is really query.search("keyword_string").  I alluded to
this at the bottom of my first post, but was hoping I could get a
generalized solution.

One approach may be to use username as the key for kind2, and thus my
query can be Kind1.all().search(keyword_string).filter
('collection_name >', key_constructed from username).filter
('collection_name <', key_constructed_from_usernameZ).fetch().  I'm
not sure it's going to work!  If there's a cleaner, more generalized
approach, I'd love to hear it!

On Feb 26, 9:58 pm, Tim Hoffman <zutes...@gmail.com> wrote:
> So in your original posting you said
>
> kind1 where prop1 = specific_value.
>
> So what specific value would you be looking for the whole text or
> specific words or combinations of words (in a particular order )
>
> This is I think where you approach will be determined.
>
> T
>
> On Feb 27, 1:41 pm, Devel63 <danstic...@gmail.com> wrote:
>
> > It's a good idea, and it's what I've used elsewhere, but in this case
> > kind1_prop is really text (I'm using the searchable data model).
> > These are not tiny, and contain data from multiple fields.  It would
> > be expensive to duplicate all this in a third class, and multiply the
> > data storage by a factor of 1 million.  So ... is there another way?
>
> > On Feb 26, 7:20 pm, Tim Hoffman <zutes...@gmail.com> wrote:
>
> > > I think you should create a new entity
> > > which has
> > >    kind1_key
> > >    kind1_prop
> > >    kind2_key
> > >    kind1_ref
> > >    kind2_ref
>
> > > Filter on kind2 key, kind 1 key , kind1 prop
> > > The use the refs to get the real objects when you need to.
>
> > > Or something like that
>
> > > T
>
> > > On Feb 27, 4:09 am, Devel63 <danstic...@gmail.com> wrote:
>
> > > > Until now, the lack of a SQL-like join capability has not been an
> > > > issue.  But...
>
> > > > I have 2 entity kinds:
> > > >     - kind1: prop1
> > > >     - kind2: username, reference_to_kind1
>
> > > > I want to find all kind2's where user=me and reference_to_kind1 is a
> > > > kind1 where prop1 = specific_value.
>
> > > > I can't figure out a way to do this except to start with either class,
> > > > do a fetch, then iterate over all references.  This is potentially A
> > > > LOT of individual queries.  The intersection/join is small, but each
> > > > kind can have large numbers of entities.
>
> > > > How is this done?
>
> > > > P.S. Please don't suggest denormalization, because the above is a
> > > > simplified example.  I am really doing this with text fields in kind1,
> > > > and keyword searchables.
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"Google App Engine" group.
To post to this group, send email to google-appengine@googlegroups.com
To unsubscribe from this group, send email to 
google-appengine+unsubscr...@googlegroups.com
For more options, visit this group at 
http://groups.google.com/group/google-appengine?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to