writeRaster: Save a GRaster to disk

In fasterRaster: Faster Raster and Spatial Vector Processing Using 'GRASS GIS'

Save a GRaster to disk

Description

This function saves a GRaster to disk directly from a GRASS session. It is faster than using rast(), then saving the output of that to disk (because rast() actually save the raster to disk, anyway).

The function will attempt to ascertain the file type to be ascertained from the file extension, but you can specify the format using the format argument (see entry for ...). You can see a list of supported formats by simply using this function with no arguments, as in writeRaster(), or by consulting the online help page for the GRASS module r.out.gdal (see grassHelp("r.out.gdal")). Only the GeoTIFF file format is guaranteed to work for multi-layered rasters.

The function will attempt to optimize the datatype argument, but this can take a long time. You can speed this up by setting datatype manually. Note that if you are saving a "stack" of GRasters with different datatypes, the one with the highest information density will be used (e.g., low-bit integer < high-bit integer < floating-point < double-floating point). This can make rasters with lower datatypes much larger on disk. In these cases, it make be best to save rasters with similar datatypes together.

Usage

## S4 method for signature 'GRaster,character'
writeRaster(
  x,
  filename,
  overwrite = FALSE,
  datatype = NULL,
  byLayer = FALSE,
  names = TRUE,
  levelsExt = NULL,
  compress = "LZW",
  warn = TRUE,
  ...
)

## S4 method for signature 'missing,missing'
writeRaster(x, filename)

Arguments

x	A GRaster or missing: If missing, a table of supported file types is reported.
filename	Character: Path and file name.
overwrite	Logical: If FALSE (default), do not save over existing file(s).
datatype	NULL (default) or character: The datatype of the values stored in non-ASCII rasters. If NULL, this will be ascertained from the raster, and the function usually does a good job at it. However, you can force it manually, but note that in some cases, trying to save a GRaster using an inappropriate datatype for its values can result in an error or in the function exiting without an error but also without having written the raster to disk. The argument can take any of those shown below under the first four columns, but whatever is used, it will be converted to the GDAL version. fasterRaster terra GRASS GDAL Values integer INT1U CELL Byte Integer values from 0 to 255 integer INT2U CELL UInt16 Integer values from 0 to 65,534 integer INT2S CELL Int16 Integer values from -32,767 to -32,767 integer INT4S CELL Int32 Integer values from -2,147,483,647 to 2,147,483,647 float FLT4S FCELL Float32 Values from -3.4E+38 to 3.4E+38, including decimal values double FLT8S DCELL Float64 Values from -1.79E+308 to 1.79E+308, including decimal values factor INT* CELL ⁠INT*⁠ Integer values corresponding to categories * Depends on the integers (signed/unsigned, range of values). Categorical rasters will have an associated file saved with them that has category values and labels. The file name will be the same as the raster's file name, but end with the extension given by levelsExt (.csv by default).
byLayer	Logical: If FALSE (default), multi-layer rasters will be saved in one file. If TRUE, the each layer will be saved in a separate file. The filename from filename will be amended so that it ends with ⁠_<name>⁠ (then the file extension), where ⁠<name>⁠ is give by names(). Note that if any characters in raster names will not work in a file name, then the function will fail (e.g., a backslash or question mark).
names	Logical: If TRUE (default), save a file with raster layer names. The file will have the same name as the raster file but end with "⁠_names.csv⁠". Currently, the names() attribute of rasters cannot be saved in the raster, which can create confusion when multi-layered rasters are saved. Turning on this option will save the ancillary file with layer names. If it exists, this file will be read by fast() so layer names are assigned when the raster is read by that function. The absence of a "names" file will not create any issues with this function or fast(), other than not having the metadata on layer names.
levelsExt	Character, logical, or NULL (default): Name of the file extension for the "levels" file that accompanies a categorical GRaster. When saving categorical rasters, the raster file is accompanied with a "levels" file that contain information on the levels of the raster. This file is the same as filename, except it has a different extension. Valid values depend on how many raster layers are saved at a time (case is ignored): DefaultOne raster layer: ".csv" Two or more layers, with at least one categorical raster: ".rds", ".rda", ".rdat", ".rdata" Any: NULL or TRUE automatically selects either ".csv" (one raster layer) or ⁠".rds⁠ (two or more) Any: FALSE disables saving of a levels file.
compress	Character: Type of compression to use for GeoTIFF files: "LZW" (default) "DEFLATE" "PACKBITS" "LZMA" NULL: No compression is used, but the file can still be reduced in size by using zip, gzip, or other compressions.
warn	Logical: If TRUE (default), display a warning if the datatype argument does not match the value given by datatype(x, "GDAL"), or if the fileExt argument will not work with the given raster and so has been automatically changed.
...	Additional arguments. These can include: bigTiff: Logical: If TRUE, and the file format is a GeoTIFF and would be larger than 4 GB (regardless of compression), then the file will be saved in BIGTIFF format. format: Character, indicating file format. This is usually ascertained from the file extension, but in case this fails, it can be stated explicitly. When using other formats, you may have to specify the createopts argument, too (see help page for GRASS module r.out.gdal). Two common formats include: "GTiff" (default): GeoTIFF filename ends in .tif. "ASC": ASCII filename ends in .asc Additional arguments to send to GRASS modules r.out.gdal and r.out.ascii. precision: Numeric: For ASCII files, you may need to state the number of significant digits. 32-bit values have 7 digits and 64-bit values have 16. So in these cases the argument would be precision=7 or precision=16.

Value

A GRaster (invisibly). A raster is also saved to disk.

See Also

terra::writeRaster(), GRASS module r.out.gdal (see grassHelp("r.out.gdal"))

Examples

if (grassStarted()) {

# Setup
library(terra)

# Example data
madElev <- fastData("madElev")
madChelsa <- fastData("madChelsa")

### What raster formats can we attempt to write?
writeRaster()

### Save GRaster to disk (using temporary file)
elev <- fast(madElev)
filename <- tempfile(fileext = ".tif")
writeRaster(elev, filename)

# Load raster from disk
elev2 <- fast(filename)
elev2

### Save multi-layer GRaster to disk in one file (using temporary file)
chelsa <- fast(madChelsa)
filename <- tempfile(fileext = ".tif")
writeRaster(chelsa, filename)

# Load raster from disk
chelsa2 <- fast(filename)
chelsa2

### Save multi-layer GRaster to disk layer-by-layer (using temporary file)
chelsa <- fast(madChelsa)
filename <- tempfile(fileext = ".tif")
writeRaster(chelsa, filename, byLayer = TRUE)

# Load one of the rasters from disk
filename2 <- sub(filename, pattern = ".tif", replacement = "_bio1.tif")
chelsaBio1 <- fast(filename2)
chelsaBio1

}

fasterRaster documentation built on June 8, 2025, 12:17 p.m.