Thanks, but unfortunately, I need to fetch multiple super columns - each
super column represents a particular "state" and as entities transition from
one state to another, they move from one super column to another.

So, there is no way currently I can retrieve multiple super columns with
their sub-cols reverse sorted? If I need it, I will need to fire a separate
query for each "state" super column?


On Mon, Dec 27, 2010 at 9:59 AM, Jonathan Ellis <jbel...@gmail.com> wrote:

> If you are fetching multiple supercolumns, then that's what you can
> control the order of.  If you are slicing within a single supercolumn,
> the reverse parameter will affect the order of subcolumns.
>
> On Sun, Dec 26, 2010 at 6:11 AM, Roshan Dawrani <roshandawr...@gmail.com>
> wrote:
> > Hi Ran,
> > I am not doing it the YAML way. I am defining my SCF through Hector API
> as
> > below
> >
> > ================================================================
> > ThriftCfDef cfDef = HFactory.createColumnFamilyDefinition(ksName, "SCF")
> > cfDef.setColumnMetadata(...)
> > cfDef.setColumnType(ColumnType.SUPER)
> > cfDef.setSubComparatorType(ComparatorType.TIMEUUIDTYPE)
> > ================================================================
> >
> > And this is the structure of the SCF:
> > ================================================================
> > "key1" :
> >           "superCol1":
> >                     SubColName1 (TimeUUID) : null
> >                     SubColName2 (TimeUUID) : null
> >                     SubColName3 (TimeUUID) : null
> >           "superCol2":
> >                     SubColName4 (TimeUUID) : null
> >                     SubColName5 (TimeUUID) : null
> >                     SubColName6 (TimeUUID) : null
> > ================================================================
> >
> > I now try to retrieve the data as below, but for both reverse = true |
> > false, it's only the order of supercolumns that changes. The subcolumns
> > always come in the same order - oldest to newest
> >
> > ================================================================
> > RangeSuperSlicesQuery#setRange (null, null, reverse, Integer.MAX_VALUE)
> //
> > reverse = true | false
> > ================================================================
> >
> > Anything I am doing wrong here?
> >
> > --
> > Roshan
> > Blog: http://roshandawrani.wordpress.com/
> > Twitter: @roshandawrani
> > Skype: roshandawrani
> >
> > On Sun, Dec 26, 2010 at 5:31 PM, Ran Tavory <ran...@gmail.com> wrote:
> >>
> >> Roshan, in cassandra.yaml did you define compare_subcolumns_with  for
> the
> >> SCF?
> >> The subcolumn names are the timeuuid, not the subcolumn values, right?
> >>
> >> On Sun, Dec 26, 2010 at 9:14 AM, Roshan Dawrani <
> roshandawr...@gmail.com>
> >> wrote:
> >>>
> >>> Hi,
> >>>
> >>> I have a super ColumnFamily that has a few super columns, with each
> >>> having a growing list of sub-columns where column name is a TimeUUID,
> so the
> >>> sub-columns get stored chronologically (oldest to newest)
> >>>
> >>> When I retrieve this data, can I somehow retrieve the sub-columns in
> >>> reverse order - newest first?
> >>>
> >>> I am using RangeSuperSlicesQuery to query the super columns and setting
> a
> >>> range on it with reverse = true, but that only sorts the data by super
> >>> column names.
> >>>
> >>> How can I tell RangeSuperSlicesQuery to get the sub-columns also in
> >>> reverse order?
> >>>
> >>> Thanks.
> >>>
> >>> --
> >>> Roshan
> >>> Blog: http://roshandawrani.wordpress.com/
> >>> Twitter: @roshandawrani
> >>> Skype: roshandawrani
> >>
> >>
> >>
> >> --
> >> /Ran
> >
> >
>
>
>
> --
> Jonathan Ellis
> Project Chair, Apache Cassandra
> co-founder of Riptano, the source for professional Cassandra support
> http://riptano.com
>

Reply via email to