Re: Remove duplicate stations

I don't know if this of interest, but we have an applications example
that shows how to take text strings that mark locations of stations
and remove the ones that overlap. See example 3 at:

   http://www.ncl.ucar.edu/Applications/station.shtml:

--Mary

On Thu, 28 Dec 2006, Dennis Shea wrote:

>> I am working with a large volume of station data thats in a absolute mess!! I
>> am talking about the GHCN Daily temperature data. There are several stations
>> that are located exactly at the lat, long coordinate and have overlapping time
>> series. I need to find all such stations and flag them or remove them. The
>> station data is arranged in a 2D grid of the form (STNID,T) where STNID is
>> unique for each station and T is days from 1 to N. Is there any way to achieve
>> the kind of flagging that I am talking about without going through tedious for
>> loops?
> ==========================================================
> Comments:
> [a] If N is of arbitrary length, it is not a 'nice' 2D array.
> [b] You say that the stations IDs are unique but you never
> mention lat/lon information.
> [c] There is no function that does what you want.
> [d] Something like the following must be used:
>
> id = station_data(:,0) ; station IDs only
> nid = dimsizes(id)
> n_dup = 0
>
> do n=0,nid-2
> i = ind(id.eq.id(n))
> if (.not.all(ismissing(i))) then
> ni = dimsizes(i)
> print("id="+id(n)+" has "+ni +" duplicates")
> ;print("i="+i)
> delete(i)
>
> n_dup = n_dup+1
> end if
> end do
> print("n_dup="+n_dup)
>
> This is far from perfect. You may wish to preallocate
> space for duplicate IDs and have more code to avoid
> duplicate printing.
>
> id_dup = new( 1000, typeof(id)) ; 1000 arbitrary
> id_dup = 0
>
> Then add code to add entries to the above array.
>
> good luck
>
> _______________________________________________
> ncl-talk mailing list
> ncl-talk_at_ucar.edu
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
_______________________________________________
ncl-talk mailing list
ncl-talk_at_ucar.edu
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Thu Dec 28 2006 - 09:38:29 MST