Change the Missing Value For a netCDF Variable
Changes the missing_value attribute for a netCDF variable.
ncvar_change_missval( nc, varid, missval )
Note: this specialty function is only used to change a variable's missing value
after it has already been defined,
which is rare. The proper way to set
a variable's missing value in the first place is by setting the missing value argument to
routine ncvar_def appropriately.
Missing values are special values in netCDF files whose value is to be taken as indicating the data is "missing". This is a convention, and is indicated by the netCDF variable having an attribute named "missing_value" that holds this number. This function sets the "missing_value" attribute for a variable.
R uses a similar concept to indicate missing values, the "NA" value. When the ncdf library reads in data set from a pre-existing file, all data values that equal that variable's missing value attribute appear to the R code as being "NA" values. When the R code writes values to a netCDF variable, any "NA" values are set to that variable's missing value before being written out. This makes the mapping between netCDF's "missing_value" attribute and R's "NA" values transparent to the user.
For this to work, though, the user still has to specify a missing value
for a variable. Usually this is specified when the variable is created,
as a required argument to ncvar_def.
However, sometimes it is useful to add (or change) a missing value for variable
that already exists in a disk file. This function enables that.
David W. Pierce dpierce@ucsd.edu
http://dwpierce.com/software
# Make an example netCDF file with a given missing value. We will # then change the missing value in the file using ncvar_change_missval origMissVal <- -1. dimX <- ncdim_def( "X", "meters", 1:7 ) varAlt <- ncvar_def( "Altitude", "km", dimX, origMissVal ) ncnew <- nc_create( "transect.nc", varAlt ) data <- c(10.,2.,NA,1.,7.,NA,8.) ncvar_put( ncnew, varAlt, data ) nc_close(ncnew) # At this point, the actual data values in the netCDF # file will be: 10 2 -1 1 7 -1 8 # because the "NA" values were filled with the missing # value, -1. Also, the missing_value attribute of variable # "varAlt" will be equal to -1. # Now change the missing value to something else. Remember # we have to open the file as writable to be able to change # the missing value on disk! newMissVal <- 999.9 nc <- nc_open( "transect.nc", write=TRUE ) varname <- "Altitude" data <- ncvar_get( nc, varname ) # data now has: 10., 2., NA, 1., 7., NA, 8. print(data) ncvar_change_missval( nc, varname, newMissVal ) ncvar_put( nc, varname, data ) nc_close(nc) # Now, the actual data values in the netCDF file will be: # 10 2 999.9 1 7 999.9 8 # and the variables "missing_value" attributre will be 999.9 # **NOTE** that we had to explicitly read in the data and write # it out again in order for the on-disk missing values in the # data array to change! The on-disk missing_value attribute for # the variable is set automatically by this function, but it is # up to you whether or not you want to read in all the existing # data and change the values to the new missing value. # Clean up our example file.remove( "transect.nc" )
Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.