The goal of ncdf4 is to provide an "Interface to Unidata netCDF (Version 4 or Earlier) Format Data Files".
This fork of the CRAN package ncdf4 is intended to update the Windows build process, and specifically to add support for Thredds.
Alternatives:
- Unidata provides Visual Studio binaries that can be used to put a build together
- pmjherman has this repo
- check Makevars.win (assume can remove "DHAVE_LIBUDUNITS2" but what about "-DHAVE_DECL_NC_RENAME_GRP=1"?)
- add tests, especially for Thredds (make sure 32-bit Windows limitation is understood)
- roxygenize?
You can install the released version of ncdf4 from CRAN with:
install.packages("ncdf4")And the development version from GitHub with:
# install.packages("devtools")
devtools::install_github("mdsumner/ncdf4")Read from Thredds (will only work in Windows 64-bit, not 32-bit).
u <- "http://coastwatch.pfeg.noaa.gov/erddap/griddap/erdQSwind3day"
nc <- ncdf4::nc_open(u)
class(nc)
#> [1] "ncdf4"
## note that "ncdf" argument is required to trigger use of ncdf4
r <- raster::raster(u, ncdf = TRUE, varname = "x_wind")
print(r)
#> class : RasterLayer
#> band : 1 (of 3774 bands)
#> dimensions : 1201, 2881, 3460081 (nrow, ncol, ncell)
#> resolution : 0.125, 0.125 (x, y)
#> extent : -0.0625, 360.0625, -75.0625, 75.0625 (xmin, xmax, ymin, ymax)
#> coord. ref. : +proj=longlat +datum=WGS84 +ellps=WGS84 +towgs84=0,0,0
#> data source : http://coastwatch.pfeg.noaa.gov/erddap/griddap/erdQSwind3day
#> names : Zonal.Wind
#> z-value : 932644800
#> zvar : x_wind
#> level : 1