Page 1 of 1

problem with quantile function

Posted: Mon May 21, 2012 2:12 am
by tospol.kaw
-9018.631165000001
-4107.883231
-2741.856
-1977.900775
-1363.424099
-695.0278893
-366.0882408
-258.1590605
-5.82e-11
-5.82e-11
-5.82e-11
-5.82e-11
-5.82e-11
-5.82e-11
-5.82e-11
-5.82e-11
-2.91e-11
-2.91e-11
-2.91e-11
-2.91e-11
-2.91e-11
-2.91e-11
-2.91e-11

According to function @quantile(xxx , 0.02 , "g"), why can't I calculate 2nd percentile of above data?
And why option of this function can't change the result in any rank?

Re: problem with quantile function

Posted: Mon May 21, 2012 7:40 am
by EViews Gareth
What do you mean you can't calculate it? I just tried it on that data, and it returned a result.

Re: problem with quantile function

Posted: Mon May 21, 2012 8:06 am
by startz
When I try it I get all NA. Could this have to do with asking for the second percentile with only 3 obs?

Re: problem with quantile function

Posted: Mon May 21, 2012 8:28 am
by EViews Gareth
Ah, it seems to depend on whether you use it as a series or a scalar expression. Scalar seems to work. I'll take a look at why series is not.

Re: problem with quantile function

Posted: Mon May 21, 2012 9:24 am
by EViews Gareth
Ok, there was a bug in the series version of @quantile that meant that the "g" option wasn't being recognised. We'll fix that in the next patch. In the meantime you can change your code to be:

Code: Select all

@quantile(xxx , 0.02 , 6)
where the "6" is a magic option that means "gumbel", and it works without the fix.

Re: problem with quantile function

Posted: Tue May 22, 2012 12:55 am
by tospol.kaw
Thanks a lot.