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

Re: skip, stride reading...

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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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: skip, stride reading...
  • Date: Fri, 17 Oct 2003 14:41:57 -0600
Tom Rink wrote:


Hi,

How can skip or stride reading be done in the Java
implementation of NetCDF? Is there an equivalent
operation?

Tom Rink


Hi Tom:
I seem to have overlooked implementing a stride-based read, probably because the original (version 1) library doesnt have one. I will add this to the to-do list.
Meanwhile, you can get a strided subset of an Array, which will be equivilent except that you probably have to read more than you need:
Suppose you have a 3D Array, and you want to access it with stride 2, 1, and 3 in the three dimensions. The following will give you a logical view of the data with those strides, using the same backing data as the original: 

Array data = var.read();
int[] shape = data.getShape();
Array dataS;
Range[] ranges = new Range[] { new Range(0,shape[0]-1, 2), null, new Range(0,shape[2]-1, 3) }; 

try {
dataS = data.section( ranges);
} catch (InvalidRangeException e) {
System.out.println("testStride failed == "+ e);
}
Note that the null second Range means “use all with stride one”. You can make a copy to get just the subset you want, so that the original data memory can be freed up: 

dataS = dataS.copy(); // now is a separate data array