reading tif file using terra and raster package gives different results-CodePudding

I am reading a raster file using both terra and raster package

library(raster)
library(terra)

fl_terra <- terra::rast('my_raster.tif')
fl_raster <- raster::raster('my_raster.tif')
  
fl_terra
class       : SpatRaster 
dimensions  : 157450, 327979, 1  (nrow, ncol, nlyr)
resolution  : 0.0002777778, 0.0002777778  (x, y)
extent      : -142.4107, -51.30541, 41.12569, 84.86181  (xmin, xmax, ymin, ymax)
coord. ref. : lon/lat WGS 84 (EPSG:4326) 
source      : xxxx.tif 
categories  : DepthBand 
name        :         DepthBand 
min value   : 0.0m <= x <= 0.3m 
max value   :          x > 9.0m 

fl_raster
class      : RasterLayer 
dimensions : 157450, 327979, 51640293550  (nrow, ncol, ncell)
resolution : 0.0002777778, 0.0002777778  (x, y)
extent     : -142.4107, -51.30541, 41.12569, 84.86181  (xmin, xmax, ymin, ymax)
crs        :  proj=longlat  datum=WGS84  no_defs 
source     : xxxx.tif 
names      : DepthBand 
values     : 1, 6  (min, max)

Why is reading the same file using the two packages showing different values? If I open the same file in QGIS, the legend displays the values present in the fl_raster i.e. using the raster package.

CodePudding user response：

The file has categorical values. fl_terra tells you that:

categories  : DepthBand

And it also properly shows the (alphabetical) range of the categories (and it also uses them in the legend when using plot, and returns them when extracting cell values).

name        :         DepthBand 
min value   : 0.0m <= x <= 0.3m 
max value   :          x > 9.0m

If you do not care about the categories, you can remove them with

levels(fl_terra) <- NULL

In contrast, fl_raster shows the numerical codes that are used to represent the categories (because the raster files can only store numeric cell values).

values     : 1, 6  (min, max)

That makes it look like you have numeric values. But that is not the case.

The situation is similar to having a factor in R

f <- as.factor(c("red", "blue", "green"))

"terra" would show the category labels

f
#[1] red   blue  green
#Levels: blue green red

Whereas "raster" would show the equivalent of

as.integer(f)
#[1] 3 1 2