[las_users] Re: [ferret_users] bug with Ferret v6 ?

[Ansley Manke <ansley.b.manke@xxxxxxxx>]

Hi Jaison and Jerome,
That's indeed what's happening. Interestingly, this does not always 
occur with variables whose missing_value flag is not in an input file; 
in this case a variable is are generally assigned the default 
missing-value flag of -1.e34. This is was something we test in our 
benchmark scripts before a Ferret release, but something's wrong which 
causes this one to get the wrong flag. I'll track this down and we'll 
get the fix into our upcoming bug-fix release of Ferret. 
Ansley

Jaison Kurian wrote:
> Hi Ansley & Jerome,
>                      It seems like Ferret_V6 assigns "0" as the
> missing value if the "variable in NetCDF file" does not have an
> explicit "missing_value" attribute. (It was "-1.E+34" with earlier
> versions of Ferret). Here the first value of trange is "0" and 
> hence it will be set to "missing value" in Ferret_V6.
>
> [jaison@mercury jerome]$ f
>         NOAA/PMEL TMAP
>         FERRET v6
>         Linux(g77) 2.4.21-32 - 08/23/06
>         30-Sep-06 01:32
>
> yes? use 259da1dac3f7750fe1fc83930e5a24b6.nc
> yes? say `trange,r=bad`
>  !-> MESSAGE/CONTINUE 0
> 0
> yes? say `r_temp,r=bad`
>  !-> MESSAGE/CONTINUE -9999
> -9999
> yes? say `numprofs,r=bad`
>  !-> MESSAGE/CONTINUE 0
> 0
> yes? say `numobs,r=bad`
>  !-> MESSAGE/CONTINUE 0
> 0
> yes?
>
> Variables "trange", "numprofs" and "numobs" does not have any 
> explicit missing_value attribute, hence Ferret_V6 assigns "0"
> as the missing value. For those variables (like r_temp) which 
> have a valid missing_value attribute and value, Ferret_V6 behaves
> properly.
>
>  This problem with missing_value appers to be there only for the
> variables readed in from a NetCDF file. If we define a new variable
> in Ferret_V6, the missing value is "-1.E+34" as with earlier versions.
>
> Hope this helps,
>
> Jaison
>
> On Fri, 29 Sep 2006, Ansley Manke wrote:
>
>   
> > Hi Jerome,
> > Thanks for sending the file. Indeed, running Ferret from the command 
> > line, with Ferret v5.81 the variable TRANGE is read correctly, but with 
> > Ferret v6.0
> >
> >             NOAA/PMEL TMAP
> >             FERRET v6 
> >             Linux(g77) 2.4.21-32 - 08/23/06
> >             29-Sep-06 11:34    
> >     yes? use 259da1dac3f7750fe1fc83930e5a24b6.nc
> >      
> >     yes? list trange
> >                  VARIABLE :   (hours)
> >                  FILENAME : 259da1dac3f7750fe1fc83930e5a24b6.nc
> >                  SUBSET   : 2 points (TIME)
> >      JAN-2001 / 1:    ....
> >      DEC-2002 / 2:  17496.
> >
> >
> > we'll track down why this is.
> >
> > Ansley
> >
> > Jerome King wrote:
> >     
> > > Hi,
> > >
> > > I am having an issue with using LAS with Ferret v6. I am not having 
> > > this problem using LAS with Ferret 5.81. Both LAS are exactly the same 
> > > (6.5.1), same xml files, same query.
> > > So I am thinking it's a bug with Ferret v6 but I am not sure.
> > >
> > > The issue happens after LAS creates a NetCDF file with the insitu 
> > > package and asks Ferret to make a plot.
> > > There seems to be an issue with Ferret reading "trange".
> > > I have attached the NetCDF file created by LAS. and I have copied 
> > > below the debug ferret part for both Ferret version.
> > > I hope this helps find out what the issue is.
> > > Hope this helps,
> > > Jerome.
> > >
> > > With Ferret 5.81, I get a nice plot:
> > > ------------------------------------------------------------------------------------------------------------------------- 
> > > yes? cancel mode verify
> > > yes? cancel mode interp
> > > yes? set mem/size=80
> > > Cached data cleared from memory
> > > yes? define symbol memsize = 80
> > > yes? canc data/all
> > > yes? go std_initialize "output/259da1dac3f7750fe1fc83930e5a24b6.nc" 
> > > "1" "1" "R_TEMP"
> > > yes? set 
> > > region/x="-126.8":"-117.1"/y="34.1":"38.0"/z="0":"0"/t="01-Jan-2001 
> > > 00:00:00":"31-Dec-2002 00:00:00"
> > > yes? GO "jnls/insitu/insitu_setup.jnl" 
> > > "output/259da1dac3f7750fe1fc83930e5a24b6.nc" "TAX" "CalCOFI Report 
> > > Data" "Temperature"
> > > yes? go insitu_meta_xy "output/259da1dac3f7750fe1fc83930e5a24b6.nc" 
> > > TAX -126.8:-117.1 34.1:38.0 0:0
> > > yes? FRAME/FORMAT=gif/FILE="output/3ccada4923f5f69e8e049265df5437b1.gif"
> > > yes? quit
> > > Waiting for program to exit...
> > > program exited...
> > > LAS::Server::Handler::run(): 0.003233 secs to perform all 
> > > 'postExecute' activities (none typically)
> > > ----------------------------------------------------------------------------------------------------------------------------------------- 
> > >
> > > With Ferret v6, I get an error:
> > > ------------------------------------------------------------------------------------------------------------------------- 
> > > yes? cancel mode verify
> > > yes? cancel mode interp
> > > yes? set mem/size=80
> > > Cached data cleared from memory
> > > yes? define symbol memsize = 80
> > > yes? canc data/all
> > > yes? go std_initialize "output/259da1dac3f7750fe1fc83930e5a24b6.nc" 
> > > "1" "1" "R_TEMP"
> > > yes? set 
> > > region/x="-126.8":"-117.1"/y="34.1":"38.0"/z="0":"0"/t="01-Jan-2001 
> > > 00:00:00":"31-Dec-2002 00:00:00"
> > > yes? GO "jnls/insitu/insitu_setup.jnl" 
> > > "output/259da1dac3f7750fe1fc83930e5a24b6.nc" "TAX" "CalCOFI Report 
> > > Data" "Temperature"
> > > **ERROR: command syntax: bad
> > > define axis/t=bad:17496/npoints=100/unit=hours/t0="01-JAN-2001 
> > > 00:00:00" tfull
> > > Command file, command group, or REPEAT execution aborted
> > >
> > > killing...Waiting for program to exit...
> > > program exited...
> > > killing...Waiting for program to exit...
> > > program exited...
> > > ------------------------------------------------------------------------------------------------------------------------------------------ 
> > >
> > >       
> >     
--

~>   ~>   ~>   ~>   ~>   ~>   ~>   ~>   ~>   ~>   ~> 
Ansley Manke, NOAA/PMEL   ansley.b.manke@noaa.gov
7600 Sand Point Way NE, Seattle WA
Phone 206-526-6246,  FAX 206-526-6744