Looks like you are trying to do group-by at top level object (Pdx); can you
organize the fields so that one of the fields can be used for
group-by...will that work for your application...
Or you can do group-by on multiple fields of the object.
E.g. select pickup_cell.x, pickup_cell.y group by (pickup_cell.x,
pickup_cell.y)
-Anil.
On Fri, Jan 29, 2016 at 9:57 AM, Darrel Schneider <dschnei...@pivotal.io>
wrote:
> I think the PDX comparable conversation you remember was about Enums.
>
> When Enums are serialized by PDX we remember their ordinal value so they
> can be compared later (Enums are Comparable). But the PdxInstance used to
> represent a non-enum domain class is not Comparable.
>
>
> On Thu, Jan 28, 2016 at 10:16 PM, William Markito <wmark...@pivotal.io>
> wrote:
>
> > It's serialized in all nodes so the hack might work... I'll try.
> >
> > But I do remember Darrel talking about making PDX comparable to solve
> some
> > problem but maybe in another context, not OQL
> >
> > Not using PDX is not an option given that I was looking for the benefit
> of
> > dealing and deserializing only the fields needed by the query.
> >
> > Thanks!
> >
> > Sent from my iPhone
> >
> > > On Jan 29, 2016, at 3:07 AM, Anilkumar Gingade <aging...@pivotal.io>
> > wrote:
> > >
> > > It works if its PdxSerialized in all node...
> > >
> > > Depending on how the values are inserted, it could be in domain object
> > form
> > > (local put); in that case getObject() won't work...
> > >
> > > Or you could write your own method that could return the Object. In
> > query
> > > you can invoke that method to get the object.
> > >
> > > -Anil.
> > >
> > >
> > >
> > >> On Thu, Jan 28, 2016 at 5:25 PM, Dan Smith <dsm...@pivotal.io> wrote:
> > >>
> > >> I wonder if you could invoke a getObject on the PdxInstanceImpl to get
> > back
> > >> the real object. Something like this maybe? Bit of a hack...
> > >>
> > >> select pickup_cell.getObject from /TaxiTrip group by
> > pickup_cell.getObject
> > >>
> > >> -Dan
> > >>
> > >>
> > >> On Thu, Jan 28, 2016 at 4:39 PM, Anilkumar Gingade <
> aging...@pivotal.io
> > >
> > >> wrote:
> > >>
> > >>> PdxInstanceImpl is not comparable...
> > >>> Options are:
> > >>> - group by on inner non-pdx objects (most of the usecases what we
> have
> > >>> seen).
> > >>> - or not using pdx
> > >>>
> > >>> -Anil.
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>> On Thu, Jan 28, 2016 at 4:33 PM, Jason Huynh <jhu...@pivotal.io>
> > wrote:
> > >>>>
> > >>>> What's probably happening is that the field is being extracted but
> > >> into a
> > >>>> pdxinstanceimpl instead of the actual object that the pdx instance
> > >>>> represents. I think we've seen similar issues in the past but I
> can't
> > >>>> remember if there was a work around or if we fixed it...
> > >>>>
> > >>>> On Thu, Jan 28, 2016 at 4:07 PM, William Markito <
> wmark...@pivotal.io
> > >
> > >>>> wrote:
> > >>>>
> > >>>>> Guys, quick ask for help here... Given the following object:
> > >>>>>
> > >>>>> public class Cell implements Comparable<Cell>, PdxSerializable {
> > >>>>>
> > >>>>> I'm now trying to perform a query with order by the field that has
> > >> this
> > >>>>> object, but I'm getting the following exception:
> > >>>>>
> > >>>>> gfsh> query --query="select pickup_cell from /TaxiTrip group by
> > >>>>> pickup_cell"
> > >>>>> Result : false
> > >>>>> startCount : 0
> > >>>>> endCount : 20
> > >>>>> Message : com.gemstone.gemfire.pdx.internal.PdxInstanceImpl
> cannot
> > >>> be
> > >>>>> cast to java.lang.Comparable
> > >>>>>
> > >>>>> Am I missing something ?
> > >>>>>
> > >>>>> The other comment about OQL on GFSH is that it's not actually
> > >> returning
> > >>>>> query results when the query project a single field... I'll open
> the
> > >>>> Jira.
> > >>>>>
> > >>>>> For example:
> > >>>>>
> > >>>>> gfsh>query --query="select pickup_cell from /TaxiTrip"
> > >>>>> Result : false
> > >>>>> startCount : 0
> > >>>>> endCount : 20
> > >>>>>
> > >>>>> NEXT_STEP_NAME : END
> > >>>>>
> > >>>>> gfsh>query --query="select pickup_cell, dropoff_cell from
> /TaxiTrip"
> > >>>>> Result : true
> > >>>>> startCount : 0
> > >>>>> endCount : 20
> > >>>>> Rows : 1000
> > >>>>>
> > >>>>> pickup_cell | dropoff_cell
> > >>>>> ------------------ | ------------------
> > >>>>> {"x":166,"y":-155} | {"x":164,"y":-156}
> > >>>>> {"x":167,"y":-154} | {"x":167,"y":-151}
> > >>>>> {"x":159,"y":-153} | {"x":168,"y":-151}
> > >>>>> {"x":157,"y":-173} | {"x":167,"y":-153}
> > >>>>> {"x":161,"y":-155} | {"x":162,"y":-153}
> > >>>>> {"x":161,"y":-155} | {"x":162,"y":-153}
> > >>
> >
>
