[ncl-talk] Errors using get_isolines and area_poly_sphere together

Mon Apr 29 14:26:08 MDT 2019

Hi All,

Thanks for all the replies to this thread. I'll try out the
*eliminate_dups.ncl* function that Dennis has provided above and let you
know if this resolves my issue.

Rick was correct that I was observing duplicate 2d lat lon pairs, and not
just in one dimension. I would still be curious to understand why this
happens sometimes in the get_isolines output, as it seems like having the
same exact same set of lat,lon pairs in a contour dataset would be
redundant.

Cheers,
Philippe

On Mon, Apr 29, 2019 at 1:09 PM Dennis Shea <shea at ucar.edu> wrote:

> *Also: get_unique_values
> <http://www.ncl.ucar.edu/Document/Functions/Built-in/get_unique_values.shtml>*
> may not return values in the order one might expect. See the end of this
> email.
>
> -----
> See attached library: *eliminate_dups.ncl*
>
> It  eliminates (removes) assorted duplicate sequences.
> For some reason these functions were never included in an NCL release.
>
> Some  test functions are included.
>
> Maybe:  *eliminate_dup_pairs_all *or *eliminate_dup_pairs_consecutive*
>
> is what Phillipe would want to use
>
> Let me know if these do/don't do the job.
>
> ===
> The following illustrates the *eliminate_dup_all*
>
>
> * %> ncl test_x.eliminate_dups.ncl*
>
> * %> ncl test_xy.eliminate_dups.ncl*
> ===
>
>   s     = (/"qqqq","1234","qqqq","9999","ert","vve","qqqq","9999"/)
>   sout  = *eliminate_dup_all*(s)
> *; outpu*t
>   sout   = (/ "qqqq","1234", "9999","ert","vve" /)
>
>   SOUT  = *eliminate_dup_all_unique(*s)     ; <== calls
> *get_unique_values*
> <http://www.ncl.ucar.edu/Document/Functions/Built-in/get_unique_values.shtml>
> *output:*
>   SOUT = (/ "1234","9999", "ert","qqqq", "vve" /)
>
>
> D
>
>
>
>
> On Mon, Apr 29, 2019 at 10:44 AM Rick Brownrigg <brownrig at ucar.edu> wrote:
>
>> Thanks to Rashed for replying here. However, I'm not sure
>> get_unique_values() will do the job, as it expects a 1D array, and what
>> Philippe needs is something to weed out duplicates in both lon *and* lat.
>> In looking around, I don't see an NCL function that will do this -- it
>> might likely involve writing a loop to do this yourself. I don't know if
>> one can simply replace any found duplicates with _FillValue and get correct
>> behavior, or if one must actually remove the duplicates from the array.
>>
>>
>>
>> On Mon, Apr 29, 2019 at 6:58 AM Rashed Mahmood <rashidcomsis at gmail.com>
>> wrote:
>>
>>> Hi Philippe,
>>> I think, it makes sense to expect two same lat or lon for a circle since
>>> at two different longitudes there could be possibility of getting two
>>> latitudes and vice versa. To get unique values, you can use the following
>>> function:
>>> For example:
>>>
>>>  y:= get_unique_values(y)
>>>
>>>
>>> https://www.ncl.ucar.edu/Document/Functions/Built-in/get_unique_values.shtml
>>>
>>> Cheers,
>>> Rashed
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> On Fri, Apr 26, 2019 at 2:51 PM Philippe Papin <pppapin at gmail.com>
>>> wrote:
>>>
>>>> Hi All,
>>>>
>>>> I've been running into an issue from the combined usage of the
>>>> "get_isolines" and "area_poly_sphere" functions. First, I'm using
>>>> get_isolines to isolate a contour that completes loop around some location
>>>> on the globe. Then I want to find the area within that contour.
>>>>
>>>> What I've noticed is that occasionally (not always) the get_isolines
>>>> function will find a contour with multiple identical lat,lon pairs for the
>>>> given contour. This is problematic for the area_poly_sphere code which
>>>> requires unique vertex points as far as I can tell.
>>>>
>>>> Is there anyway to avoid get_isolines from providing identical lat,lon
>>>> points for a unique contour? If not, is there a function I could use to
>>>> remove non-unique lat,lon pairs?
>>>>
>>>> Here is a print out snippet of the relevant get_isolines output of what
>>>> I'm trying to avoid (see index values 212,213).
>>>>
>>>> Variable: x
>>>>
>>>> Type: float
>>>>
>>>> Total Size: 2452 bytes
>>>>
>>>>             613 values
>>>>
>>>> Number of Dimensions: 1
>>>>
>>>> Dimensions and sizes: [points | 613]
>>>>
>>>> Coordinates:
>>>>
>>>> Number Of Attributes: 5
>>>>
>>>>   n_points : ( 4, 613, 29, 7, 21, 41, 38, 45, 29 )
>>>>
>>>>   start_point : ( 0, 4, 617, 646, 653, 674, 715, 753, 798 )
>>>>
>>>>   segment_count : 9
>>>>
>>>>   level : 2
>>>>
>>>>   _FillValue : 9.96921e+36
>>>>
>>>>
>>>> (211) 17.06323
>>>>
>>>> (212) 16
>>>>
>>>> (213) 16
>>>>
>>>> (214) 14.94309
>>>>
>>>> (215) 13.88631
>>>>
>>>> Variable: y
>>>>
>>>> Type: float
>>>>
>>>> Total Size: 2452 bytes
>>>>
>>>>             613 values
>>>>
>>>> Number of Dimensions: 1
>>>>
>>>> Dimensions and sizes: [points | 613]
>>>>
>>>> Coordinates:
>>>>
>>>> Number Of Attributes: 5
>>>>
>>>>   n_points : ( 4, 613, 29, 7, 21, 41, 38, 45, 29 )
>>>>
>>>>   start_point : ( 0, 4, 617, 646, 653, 674, 715, 753, 798 )
>>>>
>>>>   segment_count : 9
>>>>
>>>>   level : 2
>>>>
>>>>   _FillValue : 9.96921e+36
>>>>
>>>>
>>>> (211) 69.65783
>>>>
>>>> (212) 69.83716
>>>>
>>>> (213) 69.83716
>>>>
>>>> (214) 70.0269
>>>>
>>>> (215) 70.21735
>>>>
>>>> It seems like the code is outputting an unnecessary extra lat,lon pair
>>>> at 69.83716 N, 16E. Why does this happen, and how can I avoid this for
>>>> accurate spherical area calculations?
>>>>
>>>> Thanks,
>>>> Philippe
>>>> --
>>>>
>>>> -------------------------------------
>>>> Philippe P. Papin, Ph.D.
>>>> NRC Research Associate
>>>> Global Modeling Division
>>>> Naval Research Laboratory - Monterey, CA
>>>> http://www.atmos.albany.edu/student/ppapin/
>>>> _______________________________________________
>>>> ncl-talk mailing list
>>>> ncl-talk at ucar.edu
>>>> List instructions, subscriber options, unsubscribe:
>>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>>>
>>> _______________________________________________
>>> ncl-talk mailing list
>>> ncl-talk at ucar.edu
>>> List instructions, subscriber options, unsubscribe:
>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>>
>> _______________________________________________
>> ncl-talk mailing list
>> ncl-talk at ucar.edu
>> List instructions, subscriber options, unsubscribe:
>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>
> _______________________________________________
> ncl-talk mailing list
> ncl-talk at ucar.edu
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>

-- 

-------------------------------------
Philippe P. Papin, Ph.D.
NRC Research Associate
Global Modeling Division
Naval Research Laboratory - Monterey, CA
http://www.atmos.albany.edu/student/ppapin/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20190429/af325d9a/attachment.html>