Band of data missing in netcdf file

Hi Folks,

I've got a weird one for you.

I have a process that writes out netcdf files that typically works fine. The 
process takes an array of lat/lon values and for each element in the array it 
increments a 2D array in my netcdf file. The 2D array is essentially a counter 
for the number of lat/lon values in my initial array that map to each pixel (in 
lat/lon space) in my netcdf file. However, there is an abnormal case that shows 
up when I write a large number of times to the netcdf file (1050705 times) 
where a horizontal band of the data in the netcdf seems to be getting lost 
somewhere.

Here's what I've done to test:
Written smaller netcdf files using the exact same code with data in the missing 
zone. The data in this test case shows up fine.
Stepped through with a debugger to make sure my code is actually writing the 
values into the netcdf array, I can see the values being entered and feel 
confident that this is working.
I've viewed the data file in multiple data viewers to make sure it isn't a 
glitch in the data viewer (openev, gdal_translate (to an image))

About my setup:
I'm using netcdfUI-2.2.18.jar (was using 2.2.16, but updated before sending 
this email, to make sure).
Writing my data to a ucar.ma2.ArrayInt (and my values aren't coming near 
overflowing the 16bit limit)
Java 1.5.0_3
Windows XP


Here's a snapshot of what the problem looks like when visualized:
http://www.mbari.org/staff/achase/vdm/netcdf-wierdness.jpg
The banding I'm referring to is right in the middle of the screen. I set the 
viewing options to show pixels with any value as white and no data as black.

Any thoughts? Anybody come across a problem like this before?

Here's the gdalinfo output for the file:
Driver: netCDF/Network Common Data Format
Size is 512, 512
Coordinate System is `'
Metadata:
  NC_GLOBAL#creationDate=01/24/2007
  NC_GLOBAL#lastModified=01/24/2007
  NC_GLOBAL#description=Test File
  NC_GLOBAL#title=Test Data
  NC_GLOBAL#keywords=test
  NC_GLOBAL#Conventions=CF-1.0
Subdatasets:
  SUBDATASET_1_NAME=NETCDF:"physical-object.nc":physical-object
  SUBDATASET_1_DESC=[4001x5001] physical-object (16-bit integer)
Corner Coordinates:
Upper Left  (    0.0,    0.0)
Lower Left  (    0.0,  512.0)
Upper Right (  512.0,    0.0)
Lower Right (  512.0,  512.0)
Center      (  256.0,  256.0)

And for the variable in question:
C:\workspace\vdm\snapshot_creator>gdalinfo NETCDF:"physical-object.nc":physical-
object
Driver: netCDF/Network Common Data Format
Size is 5001, 4001
Coordinate System is `'
Origin = (-125.000000000000000,38.000999999999998)
Pixel Size = (0.000999800039992,-0.000999750062484)
Metadata:
  NC_GLOBAL#creationDate=01/24/2007
  NC_GLOBAL#lastModified=01/24/2007
  NC_GLOBAL#description=Test File
  NC_GLOBAL#title=Test Data
  NC_GLOBAL#keywords=test
  NC_GLOBAL#Conventions=CF-1.0
  physical-object#units=count
  physical-object#fillvalue=0
  physical-object#missing_value=0
  latitude#units=degrees_north (+N/-S)
  longitude#units=degrees_east (+E/-W)
Corner Coordinates:
Upper Left  (-125.0000000,  38.0010000)
Lower Left  (-125.0000000,  34.0010000)
Upper Right (-120.0000000,  38.0010000)
Lower Right (-120.0000000,  34.0010000)
Center      (-122.5000000,  36.0010000)
Band 1 Block=5001x1 Type=Int32, ColorInterp=Undefined
  NoData Value=0
  Metadata:
    NETCDF_VARNAME=physical-object

Thanks,
Andrew


Andrew C. Chase
Software Engineer
Monterey Bay Aquarium Research Institute
(831) 775-1865 :: Work
(831) 419-8566 :: Cell [NOTE: New number as of May 22, 2005]


===============================================================================
To unsubscribe netcdf-java, visit:
http://www.unidata.ucar.edu/mailing-list-delete-form.html
===============================================================================