Re: Coordinate subscript type mismatch.

Hi Dave,

Thanks for the information. I'll test my code as you suggested. I agree
for simple cases like mine, the variable type actually doesn't influence
later calculations and graphic outputs, but only make things complex. It's
great that a newer NCL will make this simple. Thank you!

Hongyan
On Wed, Oct 10, 2012 at 6:20 PM, David Brown <dbrown@ucar.edu> wrote:

> Hi Hongyan,
> Well it may not have worked because at line 207(now commented out) it
> looks like you repeated the error I told you about earlier, i.e. you
> convert an array from double to float and then assign it back to the
> existing double variable.
> But I am glad you figured out a way to get your code working.
> And yes NCL is strict about assignments (type and number of dimensions and
> dimension sizes all have to match). This is admittedly annoying sometimes
> and we are planning an alternate assignment operator to help avoid the
> issue in a future release.
> -dave
>
> On Oct 10, 2012, at 3:59 PM, H.Dang wrote:
>
> Hi Dave,
>
> The type of the variable seems to be very strict? It didn't work for me
> by any means, so I had to avoid using it, as I explained to Mary.
>
> The other similar problem in the same code is also related to the convert
> of variable types:
>
> 205 ;data = new((/2,28/),float, -999.)
> 206 data = new((/2,28/),double, -999.)
> 207 ;nox_MIPAS_ave_40_60NS = dble2flt(nox_MIPAS_ave_40_60NS)
> 208 data(0,:) = nox_MIPAS_ave
> 209 data(1,:) = nox_modelEY_ave_40_60NS
>
> Originally my code uses line 205 and line 207. The error message is also
> like "....can not be coerced to type of...". The problem is solved when I
> commented out line 205 and 207, and add line 206. It seems to me that the
> float variable can be assigned to double, but a double one has higher
> priority? It won't stay in the right hand side of "=" when the left side
> hand is a float.
>
> Thank you.
>
> Hongyan
> On Wed, Oct 10, 2012 at 1:15 PM, David Brown <dbrown@ucar.edu> wrote:
>
>> I think the problem is that you initially create the variable 'pre_MIPAS'
>> as a double with the statement:
>>
>> pre_MIPAS = plev_MIPAS(ipressure)
>>
>> In NCL once a variable has been created its type cannot be changed, so
>> the next statement:
>>
>> pre_MIPAS = dble2flt(plev_MIPAS(ipressure))
>>
>> does indeed create a float value on the right side of the assignment, but
>> then it is converted using normal type coercion back into a double for
>> assignment to a pre-existing double variable. (See
>> http://www.ncl.ucar.edu/Document/Manuals/Ref_Manual/NclDataTypes.shtml#Coercion
>> .)
>>
>> The cure, I think, is simply to eliminate the first statement:
>>
>> pre_MIPAS = plev_MIPAS(ipressure)
>>
>> so that you are creating 'pre_MIPAS' as a float in the first place.
>> -dave
>>
>>
>> On Oct 10, 2012, at 10:30 AM, Mary Haley wrote:
>>
>> > Hongyan,
>> >
>> > I don't think I've see this type of error before, so I don't fully
>> understand it. (Dave, can you help out here?)
>> >
>> > Line 168 is
>> >
>> >> 168 nox_modelEY_ave_40_60NS(ipressure) =
>> avg(vNOx({pre_MIPAS},{40:60},:))
>> >
>> > I need to see what those variables look like:
>> >
>> > printVarSummary(nox_modelEY_ave_40_60NS)
>> > printVarSummary(vNOx)
>> >
>> > My feeling is that there's an issue with the way you are subscripting
>> vNOx, so
>> > I need to see what its coordinate arrays look like.
>> >
>> > --Mary
>> >
>> > On Oct 8, 2012, at 8:37 PM, H.Dang wrote:
>> >
>> >> Dear all,
>> >>
>> >> I got the following error information in my NCL code:
>> >>
>> >> fatal:Coordinate subscript type mismatch. Subscript (0) can not be
>> coerced to type of coordinate variable
>> >> fatal:["Execute.c":7556]:Execute: Error occurred at or near line 168
>> in file obs_mdl.ncl
>> >>
>> >>
>> >> 161 do ipressure = 0,28 - 1
>> >> 162 pre_MIPAS = plev_MIPAS(ipressure)
>> >> 163 pre_MIPAS = dble2flt(plev_MIPAS(ipressure))
>> >> 164
>> >> 165 printVarSummary(pre_MIPAS)
>> >> 166 printVarSummary(ple40)
>> >> 167 ;
>> >> 168 nox_modelEY_ave_40_60NS(ipressure) =
>> avg(vNOx({pre_MIPAS},{40:60},:))
>> >> 179 end do
>> >>
>> >> The summary of the variables are:
>> >>
>> >> Variable: pre_MIPAS
>> >> Type: double
>> >> Total Size: 8 bytes
>> >> 1 values
>> >> Number of Dimensions: 1
>> >> Dimensions and sizes: [1]
>> >> Coordinates:
>> >> Number Of Attributes: 7
>> >> typeConversion_op_ncl : double converted to float
>> >> units : hPa
>> >> long_name : pressure
>> >> axis : Z
>> >> positive : down
>> >> standard_name : air_pressure
>> >> plev : 300
>> >>
>> >> Variable: ple40
>> >> Type: float
>> >> Total Size: 160 bytes
>> >> 40 values
>> >> Number of Dimensions: 1
>> >> Dimensions and sizes: [ple | 40]
>> >> Coordinates:
>> >> ple: [964..0.1]
>> >> Number Of Attributes: 1
>> >> units : mb
>> >>
>> >> I though the error is caused by the variable type, but when I used
>> "dble2flt" to convert the type from double to float, the error message is
>> till the same. So is my idea of converting correct or not? How should I
>> fix this problem? Thank you!
>> >>
>> >> Hongyan
>> >>
>> >> _______________________________________________
>> >> 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
>
>
>

_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Wed Oct 10 19:26:16 2012