Re: 10th and 90th percentile

stat_dispersion
http://www.ncl.ucar.edu/Document/Functions/Contributed/stat_dispersion.shtml

-Jonathan

On Apr 14, 2014, at 8:23 AM, Marston Johnston <shejo284@gmail.com> wrote:

> Here's an example that I made for my own use. It uses a function designed to keep only non-missing values, which is simple to make.
>
> undef("Percentile")
> function Percentile(x:numeric,P:float)
> begin
> if(P.gt.0) then
> P = P/100.
> end if
> x@_FillValue = default_fillvalue(typeof(x))
> print("Rain rates for project greater than 0 and less than 1000...")
> x = where((x.gt.0 .and. x.lt.1000.),x,x@_FillValue)
> nx = Crop(ndtooned(x))
> index = round(P*dimsizes(nx),3)
> print("Index of percentile: "+index+ " of size: "+dimsizes(nx))
> ib = dim_pqsort(nx,2)
> print("The "+P*100+" percentile: "+nx(index))
> return(nx(index))
> end
>
> Hope this helps,
> /M
>
>
> On Mon, Apr 14, 2014 at 2:16 PM, Ahmad Farsyud <a.farsyud@gmail.com> wrote:
> Hi Marston, thanks,
> could you suggest me an example?
>
>
> On Mon, Apr 14, 2014 at 2:13 PM, Marston Johnston <shejo284@gmail.com> wrote:
> Hi,
>
> I've not seen such a function in NCL, but it's fairly simple to code in NCL.
> It involves sorting your array, make it 1D first, removing any invalid data points, and then
> calculating the 10th and 90th percentile position in the array. The 10th and 90th are positions of the
> 10% and 90% indices in the array.
>
> Hope this helps,
>
> /M
>
>
> On Mon, Apr 14, 2014 at 2:03 PM, Ahmad Farsyud <a.farsyud@gmail.com> wrote:
> Dear NCL,
>
> Anyone can suggest what a function that i can use to calculate 10th and 90th percentile in NCL?
>
> Thanks
> AF
>
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
>
>
>
> --
> Only the fruitful thing is true!
>
>
>
>
> --
> Only the fruitful thing is true!
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk

_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Mon Apr 14 06:25:09 2014