Hi Bill,
This is the first "buglet" we've received!
The ut_calendar function is a straight interface to the Udunits  
"utCalendar" interface, which is part of the Udunits software package.
It looks like C program version of utCalendar is producing the same  
results, so NCL is behaving as expected.
However, I'm interested in finding out more about what this extra "t"  
means. I think IDV may be handling this type of units specially, or  
else they have modified the utCalendar to recognize it.
Does IDV have any documentation on this?
--Mary
On Nov 5, 2009, at 9:51 AM, Capehart, William J wrote:
>
> Hi guys,
>
> I may have found a small buglet with ut_calendar
>
> for the following time variable (and most importantly the unit  
> description)
>
> Variable: time
> Type: integer
> Total Size: 36 bytes
>             9 values
> Number of Dimensions: 1
> Dimensions and sizes:   [nt_time | 9]
> Coordinates:
> Number Of Attributes: 5
>   long_name :   forecast time
>   units :       hour since 2009-10-29t06:00:00z
>   _CoordinateAxisType : time
>   valid_max :   24
>   valid_min :   0
> (0)     0
> (1)     3
> (2)     6
> (3)     9
> (4)     12
> (5)     15
> (6)     18
> (7)     21
> (8)     24
>
> Note the t separating the Date and Time in the time@units attribute.
>
> The output of ut_calendar becomes
>
> ncl 19>  print(ut_calendar(time,3))
> Variable: unnamed (return)
> Type: double
> Total Size: 72 bytes
>             9 values
> Number of Dimensions: 1
> Dimensions and sizes:   [9]
> Coordinates:
> Number Of Attributes: 1
>   calendar :    standard
> (0)     2009102900
> (1)     2009102903
> (2)     2009102906
> (3)     2009102909
> (4)     2009102912
> (5)     2009102915
> (6)     2009102918
> (7)     2009102921
> (8)     2009103000
>
>
> when I take out the character separating the date from the time  
> (replacing it with a space), it’s ok. but as I understand it and  
> have seen in other packages (e.g., IDV), the t is accepted – in fact  
> IDV adds the “t” when it converts a GRIB file into netcdf.
>
> Can that be patched for 5.1.2?
>
> Thanks Much
> Bill
>
> _______________________________________________
> 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 Thu Nov  5 21:34:45 2009
This archive was generated by hypermail 2.1.8 : Wed Nov 11 2009 - 12:48:56 MST