[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Max length on attributes in cdl files?

This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.

  • Subject: Re: Max length on attributes in cdl files?
  • Date: Tue, 20 Aug 1996 14:41:10 -0600
> To: address@hidden
> From: address@hidden (Bill Moninger)

Hi Bill,

I just got back from vacation and read your posting:

> In an effort to maximize the readability of my cdl files, I entered
> extensive multi-line comments in the :units attribute.  This worked fine on
> a sun running SunOS, but not on HP-UX or SGI, which bumped up against a 200
> character limit.
> 
> So my question is, what is the best way to include documention like the
> following in a cdl file?
> 
> THIS FAILS ON SOME MACHINES:
> byte correctedFlag(recNum);
>  correctedFlag:long_name ="Corrected data indicator";
>  correctedFlag:units = "
>   r = raw data,
>   l = data that has  undergone lat/lon correction
>      (other than interpolation),
>   T = data that has undergone temperature correction, 
>   f = longitude and wind direction flipped, 
>   t = obs time has been set to the report receipt time
> ";

I've been trying to reproduce the problem, but was unable to on IRIX
6.2, though I did see a problem on HP-UX 9.05 that seems to be a
limitation of "lex" for which there is a work-around.

I first created the following file, named "moninger.cdl":

-------------------- moninger.cdl ------------------------
netcdf moninger {
dimensions:
        recNum = 1;
variables:
        byte correctedFlag(recNum);
 correctedFlag:long_name ="Corrected data indicator";
 correctedFlag:units = "
  r = raw data,
  l = data that has  undergone lat/lon correction
     (other than interpolation),
  T = data that has undergone temperature correction, 
  f = longitude and wind direction flipped, 
  t = obs time has been set to the report receipt time
";
}
-------------------- end of moninger.cdl ------------------------

Then I tried invoking 

    ncgen -b moninger.cdl 

on this file, using the latest version of the netCDF software (2.4.3).
On the SGI, the netCDF file "moninger.nc" was created without error.
Using ncdump on it, I get:

-----------------------------------------------------------------
netcdf moninger {
dimensions:
        recNum = 1 ;
variables:
        byte correctedFlag(recNum) ;
                correctedFlag:long_name = "Corrected data indicator" ;
                correctedFlag:units = "\n",
    "  r = raw data,\n",
    "  l = data that has  undergone lat/lon correction\n",
    "     (other than interpolation),\n",
    "  T = data that has undergone temperature correction, \n",
    "  f = longitude and wind direction flipped, \n",
    "  t = obs time has been set to the report receipt time\n",
    "" ;
 
data:
 
 correctedFlag = 129 ;
}
-----------------------------------------------------------------

which is another way to express multi-line text attributes, by
explicitly including the embedded new-lines ("\n").

Invoking 

   ncgen -b moninger.cdl

on HP-UX 9.05 yields

   ncgen: moninger.cdl line 7: out of memory

which I believe is a limitation of lex documented in the HP lex man
page:

    The token buffer in the program built by lex is of fixed length,

         yytext[YYLMAX]

    where YYLMAX is defined to be 200 unsigned characters or 400 unsigned
    characters if -m has been specified and LC_CTYPE indicates a multi-
    byte character set.  Overflow of this array is not detected in the
    lex.yy.c program.
 
If you instead enter the multi-line attributes using trailing embedded
"\n" characters, as in the output of ncdump above, things work fine and
ncgen is able to parse the resulting CDL file, even on HP-UX.

--Russ