I was just trying the verbs on a larger data set without
thinking too hard about its exact spec.  1000#data isn't
very interesting!  I agree "bigdata" might not be
representative of John's application.  Perhaps I've
misunderstood the problem.

"testa" (ie "test" with "slicea") seems ok for the
few times I've tried it.

Mike

On 05/05/2011 17:59, Raul Miller wrote:
> On Thu, May 5, 2011 at 12:40 PM, Mike Day<[email protected]>  wrote:
>>     ts'#slice bigdata'
>> 0.00981431 1.75866e6
>>     ts'#slicea bigdata'
>> 0.00385224 1.07226e6
> It's perhaps worth noting here that bigdata is quite sparse.
>
> Is this an accurate reflection of the problem space?
>
> P.S.   And test seems to claim some implementations are invalid, for
> reasons that I have not looked into.

----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to