I have been using Ferret for some time now, and I have often wished for a method of inputting sparse data sets more efficiently. What I would really like to be able to do is input data from a file where each record is an n-tuple: [I,J,K,L,V1,V2,...,Vm] Where (I,J,K,L) are the 4-dimensional matrix indices, and V1,...,Vm are values for variables. Data at all other indices would be considered missing. Unless I'm mistaken, this is not possible in Ferret at present. Is there any hope of adding this capability to version 5? A flexible input method that would allow one to read ASCII or binary records; to specify the byte-lengths of the indices and variables for binary files; and perhaps to have a scale and offset applied (like NetCDF files)would be a valuable addition to Ferret.
Attachment:
Donald S Dunbar.vcf
Description: Binary data