Re: removing zero dimensions ...

Just to answer Ufuk's original question: we don't have any function that removes the "zero dimension" or the "degenerate dimension" as we call it.

This question was brought up last year, so here's the answer (and a function) for that:

http://www.ncl.ucar.edu/Support/talk_archives/2010/0973.html

Chad, thanks for contributing your "squeeze" function.

--Mary

On Jun 1, 2011, at 12:04 PM, Chad Herman wrote:

> On 06/01/2011 02:00 PM, ncl-talk-request@ucar.edu wrote:
>> Message: 6
>> Date: Wed, 01 Jun 2011 11:48:04 +0200
>> From: "Ufuk Utku Turuncoglu (BE)"<u.utku.turuncoglu@be.itu.edu.tr>
>> Subject: removing zero dimensions ...
>> To: NCL USERS<ncl-talk@ucar.edu>
>> Message-ID:<4DE60AD4.2010102@be.itu.edu.tr>
>> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>>
>> Hi,
>>
>> I just wonder that is there any NCL function or procedure (i could not
>> find in the documentation) that removes the dimensions that it's actual
>> length is one (i call it as zero dimension). For example, i have a 4d
>> array and its dimension sizes are "tmp(10, 3, 1, 45)" and i want to
>> remove the third dimension and the result array would be "tmp(10,3,45)".
>> I actually wrote the following function to do that but i just want to be
>> sure it is a generic solution or not
>>
>> ;---------------------------------------------------------
>> ; Removes the zero dimensions from the variable
>> ; Input:
>> ; data - original data
>> ; Output:
>> ; tmp - modified data
>> ;---------------------------------------------------------
>> undef("rm_empty_dims")
>> function rm_empty_dims(data:numeric)
>> begin
>> ;--- get dimension size of the data ---
>> dims = dimsizes(data)
>> rank1 = dimsizes(dims)
>>
>> ;--- find the indices of the non-zero dimensions ---
>> indx = ind(dims .ne. 1)
>> rank2 = dimsizes(indx)
>>
>> ;--- if zero dimension exist, create new data array ---
>> if (rank1 .ne. rank2) then
>> tmp = new(dims(indx), typeof(data))
>> tmp = data
>> return(tmp)
>> else
>> return(data)
>> end if
>> end
>>
>> because, as you can see from the code i could do nothing about
>> attributes. How can be sure that result field will have correct
>> metadata? May be additional code can be implemented to ensure that. Is
>> the "tmp = data" operation safe between variables that are in different
>> rank (there could be shifting if it is not safe). Also note that the
>> current version of the code is working in my case.
>>
>> --ufuk
>
> Ufuk,
>
> I've written a function ("squeeze") to drop redundant dimensions and retain meta data. (See attached).
>
> Chad
> <squeeze.ncl>_______________________________________________
> 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 Wed Jun 1 12:30:48 2011