theres not much else to do to speed up the read for now. we will rewrite the
library using the java.nio package sometime in the future; preliminary tests
suggest we will get a factor of 2 speedup.
you can extract the byte array through:
byte[] data = (byte[]) fullA.getStorage();
William Ruddick wrote:
John / Support -
I have a huge 5120x5120 byte array representing a Sataelite image, that
I am reading out of a NetCDF file using the
NetCDF Java Libraries v.2.
It is taking me an average of 10 seconds on a P2 233 Windows 200 machine
with 512M ram.
I believe this is taking twice as long as it should be. Do you have any
suggestions?
My code is a follows:
//////////////////////////////////
fullimagebytes = new byte[(int)5120*5120];
ucar.nc2.Variable lVar = cdfFile.findVariable("image");
ucar.ma2.Array fullA = lVar.read(); // This statement takes on average
5 seconds
ucar.ma2.IndexIterator iter = fullA.getIndexIteratorFast();
int xi = 0;
while (iter.hasNext()) { // This loop takes another 5 seconds on average
fullimagebytes[xi] = iter.getByteNext();
xi++;
}
//////////////////////////////////
I believe there should be some way to go directly from a
ucar.nc2.Variable to a primitive byte array in 5 seconds or less.
Thanks,
NOAA International Division - FX-Net
-William Ruddick
ruddick@xxxxxxxxxxxx <mailto:ruddick@xxxxxxxxxxxx>
303-497-6757
