Changing the names of a RasterStack

Question

I'm trying to change the names of a RasterStack but the result is not changed. Below are the code lines with comments:

So, here's a simple loading of my .tif files

#Loading predictors' extent/mask
preds <- list.files("output/", pattern = ".tif", full.names = TRUE)
preds
 [1] "output/CHELSA_bio01_1981-2010_V.2.1.tif"           "output/CHELSA_bio02_1981-2010_V.2.1.tif"          
 [3] "output/CHELSA_bio03_1981-2010_V.2.1.tif"           "output/CHELSA_bio04_1981-2010_V.2.1.tif"          
 [5] "output/CHELSA_bio05_1981-2010_V.2.1.tif"           "output/CHELSA_bio06_1981-2010_V.2.1.tif"          
 [7] "output/CHELSA_bio07_1981-2010_V.2.1.tif"           "output/CHELSA_bio08_1981-2010_V.2.1.tif"          
 [9] "output/CHELSA_bio09_1981-2010_V.2.1.tif"           "output/CHELSA_bio10_1981-2010_V.2.1.tif"          
[11] "output/CHELSA_bio11_1981-2010_V.2.1.tif"           "output/CHELSA_bio12_1981-2010_V.2.1.tif"          
[13] "output/CHELSA_bio13_1981-2010_V.2.1.tif"           "output/CHELSA_bio14_1981-2010_V.2.1.tif"          
[15] "output/CHELSA_bio15_1981-2010_V.2.1.tif"           "output/CHELSA_bio16_1981-2010_V.2.1.tif"          
[17] "output/CHELSA_bio17_1981-2010_V.2.1.tif"           "output/CHELSA_bio18_1981-2010_V.2.1.tif"          
[19] "output/CHELSA_bio19_1981-2010_V.2.1.tif"           "output/CHELSA_cmi_mean_1981-2010_V.2.1.tif"       
[21] "output/CHELSA_hurs_mean_1981-2010_V.2.1.tif"       "output/CHELSA_npp_1981-2010_V.2.1.tif"            
[23] "output/CHELSA_pet_penman_mean_1981-2010_V.2.1.tif"

Then, I stack them

preds <- stack(preds)
names(preds)
 [1] "CHELSA_bio01_1981.2010_V.2.1"           "CHELSA_bio02_1981.2010_V.2.1"           "CHELSA_bio03_1981.2010_V.2.1"          
 [4] "CHELSA_bio04_1981.2010_V.2.1"           "CHELSA_bio05_1981.2010_V.2.1"           "CHELSA_bio06_1981.2010_V.2.1"          
 [7] "CHELSA_bio07_1981.2010_V.2.1"           "CHELSA_bio08_1981.2010_V.2.1"           "CHELSA_bio09_1981.2010_V.2.1"          
[10] "CHELSA_bio10_1981.2010_V.2.1"           "CHELSA_bio11_1981.2010_V.2.1"           "CHELSA_bio12_1981.2010_V.2.1"          
[13] "CHELSA_bio13_1981.2010_V.2.1"           "CHELSA_bio14_1981.2010_V.2.1"           "CHELSA_bio15_1981.2010_V.2.1"          
[16] "CHELSA_bio16_1981.2010_V.2.1"           "CHELSA_bio17_1981.2010_V.2.1"           "CHELSA_bio18_1981.2010_V.2.1"          
[19] "CHELSA_bio19_1981.2010_V.2.1"           "CHELSA_cmi_mean_1981.2010_V.2.1"        "CHELSA_hurs_mean_1981.2010_V.2.1"      
[22] "CHELSA_npp_1981.2010_V.2.1"             "CHELSA_pet_penman_mean_1981.2010_V.2.1"

Now I substitute the 1981.2010 with 1981-2010

preds2 <- gsub("1981.2010", "1981-2010", names(preds))
preds2
 [1] "CHELSA_bio01_1981-2010_V.2.1"           "CHELSA_bio02_1981-2010_V.2.1"           "CHELSA_bio03_1981-2010_V.2.1"          
 [4] "CHELSA_bio04_1981-2010_V.2.1"           "CHELSA_bio05_1981-2010_V.2.1"           "CHELSA_bio06_1981-2010_V.2.1"          
 [7] "CHELSA_bio07_1981-2010_V.2.1"           "CHELSA_bio08_1981-2010_V.2.1"           "CHELSA_bio09_1981-2010_V.2.1"          
[10] "CHELSA_bio10_1981-2010_V.2.1"           "CHELSA_bio11_1981-2010_V.2.1"           "CHELSA_bio12_1981-2010_V.2.1"          
[13] "CHELSA_bio13_1981-2010_V.2.1"           "CHELSA_bio14_1981-2010_V.2.1"           "CHELSA_bio15_1981-2010_V.2.1"          
[16] "CHELSA_bio16_1981-2010_V.2.1"           "CHELSA_bio17_1981-2010_V.2.1"           "CHELSA_bio18_1981-2010_V.2.1"          
[19] "CHELSA_bio19_1981-2010_V.2.1"           "CHELSA_cmi_mean_1981-2010_V.2.1"        "CHELSA_hurs_mean_1981-2010_V.2.1"      
[22] "CHELSA_npp_1981-2010_V.2.1"             "CHELSA_pet_penman_mean_1981-2010_V.2.1"

Finally, I assign the new name vector to preds, but the names remain wrong

names(preds) <- preds2
names(preds)
 [1] "CHELSA_bio01_1981.2010_V.2.1"           "CHELSA_bio02_1981.2010_V.2.1"           "CHELSA_bio03_1981.2010_V.2.1"          
 [4] "CHELSA_bio04_1981.2010_V.2.1"           "CHELSA_bio05_1981.2010_V.2.1"           "CHELSA_bio06_1981.2010_V.2.1"          
 [7] "CHELSA_bio07_1981.2010_V.2.1"           "CHELSA_bio08_1981.2010_V.2.1"           "CHELSA_bio09_1981.2010_V.2.1"          
[10] "CHELSA_bio10_1981.2010_V.2.1"           "CHELSA_bio11_1981.2010_V.2.1"           "CHELSA_bio12_1981.2010_V.2.1"          
[13] "CHELSA_bio13_1981.2010_V.2.1"           "CHELSA_bio14_1981.2010_V.2.1"           "CHELSA_bio15_1981.2010_V.2.1"          
[16] "CHELSA_bio16_1981.2010_V.2.1"           "CHELSA_bio17_1981.2010_V.2.1"           "CHELSA_bio18_1981.2010_V.2.1"          
[19] "CHELSA_bio19_1981.2010_V.2.1"           "CHELSA_cmi_mean_1981.2010_V.2.1"        "CHELSA_hurs_mean_1981.2010_V.2.1"      
[22] "CHELSA_npp_1981.2010_V.2.1"

What am I doing wrong here?

EDIT: Also, when I stack them, the hyphen changes to a point.

Answer 1

Raster, and by extension stack, names can't have minus signs in them.

> r = raster(matrix(1:12,3,4))
> names(r)
[1] "layer"
> names(r) = "the-layer"
> names(r)
[1] "the.layer"

Interestingly you can try setting the names element directly but even though the stored property has a minus sign, the printed value from the names call shows a dot:

> r@data@names = "the-layer"
> r@data@names
[1] "the-layer"
> names(r)
[1] "the.layer"

I can't see anything in ?raster::names that details limitations on the names of rasters, but I suspect the function validNames is involved.

help(raster::validNames)

Create valid names

Description:

     Create a set of valid names (trimmed, no duplicates, not starting
     with a number).

If you switch to the newer terra package, this restriction is lifted:

> tr = terra::rast(matrix(1:12,3,4))
> names(tr)
[1] "lyr.1"

Dots by default. Can we have minus signs?

> names(tr) = "lyr-1"
> names(tr)
[1] "lyr-1"

Yes!