Re: Size of dimension relating to file input being picked up as having different sizes on LHS & RHS of function

That is great thanks Dennis it solved the error. Can I ask another ncl-talk question on something you mentioned about ncl_join|20 being averaged?

Does that mean that the data from the 20 files is averaged, so that a value of say 40 of 'x day cold spell's that I'm plotting is actually 800 occurances for the 20 files giving an average of 40 occurances returned?

If so is this also true when it combines the ensemble runs-or is there an easy way of me checking something like that. It would explain why my output values are lower then I expected.

Kind regards

Dominique

________________________________________
From: Dennis Shea [shea@ucar.edu]
Sent: 13 July 2012 14:46
To: Cannavina, Dominique
Cc: ncl-talk@ucar.edu
Subject: Re: Size of dimension relating to file input being picked up as having different sizes on LHS & RHS of function

Sorry, hit 'Send' too fast

I think you want

    dim_sum_n(cold_uk1(:,{spell(nr)},:,:,:),1)

A *subtle* issue:
essentially, NCL creates a temporary variable inside the function

    temp = cold_uk1(:,{spell(nr)},:,:,:) ;==> (20,9,18,30)

so the averaging is done over the 1-th dimension of the temporary
variable

    dim_sum_n(temp,1)

On 7/13/12 7:40 AM, Dennis Shea wrote:
> The following is a guess ...
>
> xxx = dim_sum_n(cold_uk1(:,{spell(nr)},:,:,:),0)
> printVarSummary(xxx)
>
> will yield something like
>
> Variable: xxx
> > Number of Dimensions: 3
> > Dimensions and sizes: [9] x [18] x [30]
>
> The ncl_join|20 is averaged so it becomes degenerate (size=1)
> so it 'disappears'; the {spell(nr)} is size=1 (ie, degenerate)
> soi it is eliminated.
>
> So, dim0 on the left is 20 and on the right is 9,
> hence, the error message
> ---
>
> Variable: spell_tot_uk
> > Number of Dimensions: 4
> > Dimensions and sizes: [20] x [spell | 60] x [latitude | 18] x
> [longitude | 30]
> ---
> Variable: cold_uk1
> > Number of Dimensions:
> > Dimensions and sizes: [ncl_join | 20] x [run_length | 180] x
> [number | 9] x [latitude | 18] x [longitude | 30]
>
> On 7/13/12 2:51 AM, Cannavina, Dominique wrote:
>> Hi NCL Talk
>>
>> Can I ask for some help with the following error message-
>>
>> When I run the script (attached along with summary of all variables in
>> script) I get the following;
>>
>> fatal:Dimension size mismatch on subscript #0, left-hand and right-hand
>> side dimensions do not match
>>
>> fatal:["Execute.c":7556]:Execute: Error occurred at or near line 207
>>
>> Line 207 is
>>
>> spell_tot_uk(:,nr,:,:) = dim_sum_n(cold_uk1(:,{spell(nr)},:,:,:),0)
>>
>> The two variables it is picking up are given below but I can't see why
>> they are being picked up as being different sizes as dimension #0 has a
>> size of '20' for both. It relates to the number of files that have been
>> inputted.
>>
>> Variable: spell_tot_uk
>>
>> Type: float
>>
>> Total Size: 2592000 bytes
>>
>> 648000 values
>>
>> Number of Dimensions: 4
>>
>> Dimensions and sizes: [20] x [spell | 60] x [latitude | 18] x [longitude
>> | 30]
>>
>> Coordinates:
>>
>> spell: [1..60]
>>
>> latitude: [75..32.5]
>>
>> longitude: [-27.5..45]
>>
>> Number Of Attributes: 2
>>
>> spell : 1
>>
>> _FillValue : 9.96921e+36
>>
>> Variable: cold_uk1
>>
>> Type: integer
>>
>> Total Size: 69984000 bytes
>>
>> 17496000 values
>>
>> Number of Dimensions: 5
>>
>> Dimensions and sizes: [ncl_join | 20] x [run_length | 180] x [number |
>> 9] x [latitude | 18] x [longitude | 30]
>>
>> Coordinates:
>>
>> run_length: [1..180]
>>
>> number: [0..8]
>>
>> latitude: [75..32.5]
>>
>> longitude: [-27.5..45]
>>
>> Number Of Attributes: 2
>>
>> cold_crit_values : 0
>>
>> long_name : unique cold runs
>>
>> Any help much appreciated
>>
>> Kind regards
>>
>> Dominique
>>
>>
>>
>> _______________________________________________
>> 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
_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Fri Jul 13 09:42:52 2012