[Next] [Prev] [Contents] [Top]

An Introduction to Atmospheric and Oceanographic Datasets

Appendix B

Features of Common Scientific Data Formats; Access Information

Table B.1
Comparison of three scientific-data-management systems
Feature CDF HDF netCDF
Languages supported C, f77 C, f77, f90, C++ C, f77, f90, C++
Inherent data types char, short, byte, short, byte, char,
  int, float, long, float, short, long,
  double, double float,
  string   double
User-definable types no yes no
Data-conversion method XDR, native XDR, native XDR
Maximum array dimension 10 unlimited 32
Extended array dimension yes yes yes
Hyperset access yes yes yes
User-definable attributes yes yes yes
Attribute types any any any
Named dimensions no yes yes
Array-index ordering row, column row, column row, column
Shareability yes yes yes
Compression no yes no
Supporting tools yes many ncdump, ncgen, a few others
Adapted from: "Software for Portable Scientific Data Management," 1993: Brown, et al., Computers in Physics, 7, 304-308

GRIB is not a "tool" for scientific data management. It is designed for efficient archival and transmission of two dimensional gridded arrays. It is a flat file format which is "quasi-self-describing" ( a table look-up procedure is used). Access would be sequential. It is used for archival by the world's largest operational meteorological centers (NMC and ECMWF).

Table B.2
More CDF, netCDF, HDF, GRIB information
Format ftp URL
CDF nssdca.gsfc.nasa.gov
cd pub/cdf
netCDF unidata.ucar.edu
cd pub/netcdf
HDF ftp.ncsa.uiuc.edu
cd HDF
GRIB ncardata.ucar.edu
cd libraries/grib

An Introduction to Atmospheric and Oceanographic Datasets
[Next] [Prev] [Contents] [Top]