Subset a raster using row/column index

Question

When subsetting a matrix or DF, it is possible to reference row columns, such as df1[1:5, 3:10], or df3[2:4, ].

Is there any way to do this with a raster? That is, can I clip just rows 500:700, for example from a raster object?

I have tried using rasterFromCells(), but it doesn't give me the result I want (and it seems like there should be a more simple solution given R's other slick subsetting methods).

Example:

r <- raster(ncols = 50, nrow = 50)
r[] <- 1:ncell(r)

# I would like to subset the bottom 50 rows of cells, but keep it as a raster.
# However, this returns a numeric object.
rSub <- r[30:50, 1:50]

Thanks!

Do you want the cell values themselves or another raster object? — JmeCS, Mar 22 '18 at 16:57
My apologies; I have added a small example. I would like to keep it as a raster object. — gvan, Mar 22 '18 at 17:06
It is not enough clear for me. With this r[30:50, 1:50] you are subsetting the last 20 rows, right? — Seymour, Mar 22 '18 at 17:39

Seymour · Accepted Answer · 2018-03-22T17:49:57.700

10

I don't find the question very clear.

However, is this what you are looking for?

subR <- crop(r, extent(r, 30, 50, 1, 50))

plot(subR)

The function crop() of raster package does the trick because allow you to subset the raster object using rows and columns.

edited Mar 22 '18 at 17:49

answered Mar 22 '18 at 17:44

Seymour

3,104
2
22
46

This is exactly what I was looking for. When I originally read it in the Raster documentation, I took "extent" to refer to spatial extent. Sorry for the question quality. Thanks again. – gvan Mar 22 '18 at 18:03
I am pretty sure you can `crop` also using the spatial coordinate. – Seymour Mar 22 '18 at 18:06
1

And note that `extent(r, 30, 50, 1, 50))` returns a spatial extent. – Robert Hijmans Mar 23 '18 at 04:31
@gvan if this is the answer to your question you should validate it – Seymour Mar 23 '18 at 21:14

Robert Hijmans · Answer 2 · 2019-07-14T15:11:54.913

I prefer using crop as shown by Seymour. There is another way, using drop=FALSE

library(raster)
r <- raster(ncols = 10, nrow = 10)
values(r) <- 1:ncell(r)

rSub <- r[3:5, 2:3, drop=FALSE] 
rSub
#class       : RasterLayer 
#dimensions  : 3, 2, 6  (nrow, ncol, ncell)
#resolution  : 36, 18  (x, y)
#extent      : -144, -72, 0, 54  (xmin, xmax, ymin, ymax)
#coord. ref. : +proj=longlat +datum=WGS84 +ellps=WGS84 +towgs84=0,0,0 
#source      : memory
#names       : layer 
#values      : 22, 43  (min, max)

Subset a raster using row/column index

2 Answers2

Linked