Could you clarify... is this for pairs of rows, or is it n rows with n
columns, and is n a constant known before the query executes or based on
the presence of non-NULL column values?
And is this always adjacent rows using a clustering key - as opposed to a
partition key which does not guarantee adjacency?
-- Jack Krupansky
On Sat, Jan 30, 2016 at 10:28 PM, Jan wrote:
> Folks;
>
> Need some advice.
> We have a time-series application that needs the data being returned from
> C* to be flipped from the typical column based data to be row based.
>
> example :
> C*data : A B C
> D E F
>
> need returned data to be :
>A D
> B E
> C F
>
>
> Any input would be much appreciated.
>
> thanks,
> Jan
>