Apply a custom coordinate transform to an sf/sfc object (POINT/LINESTRING/POLYGON/MULTIPOLYGON)

st_transform_custom() walks through each geometry in an sf/sfc object, extracts its XY coordinates, applies a user-supplied transformation function to those coordinates, and rebuilds the geometry. It preserves the input CRS on the resulting sfc column. Polygon rings are re-closed after transformation so the first and last vertex match.

Usage

st_transform_custom(sf_obj, transform_fun, args)

Arguments

sf_obj: An object of class sf or sfc. Supported geometry types: POINT, LINESTRING, POLYGON, and MULTIPOLYGON.
transform_fun: A function that accepts a numeric matrix of coordinates with two columns (X, Y) and returns a transformed numeric matrix with the same number of rows and two columns. For example: function(coords, ...) cbind(f(coords[,1], ...), g(coords[,2], ...)).
args: A named list of additional arguments to pass to transform_fun. These are appended after the coords matrix via do.call(), i.e. do.call(transform_fun, c(list(coords), args)).

Value

An object of the same top-level class as sf_obj (sf or sfc), with the same column structure (if sf) and the same CRS as the input. Geometry coordinates are replaced by the coordinates returned by transform_fun.

Details

For POLYGON/MULTIPOLYGON, the function uses the ring indices returned by sf::st_coordinates() (L1 for rings and L2 for parts) to transform each ring independently, and then ensures each ring is explicitly closed (last vertex equals first vertex).

Error handling is per-geometry: if a geometry fails to transform, a warning is emitted and an empty geometry of the same "polygonal family" is returned to keep list lengths consistent.

The function does not modify or interpret the CRS numerically; it simply preserves the CRS attribute on the output sfc. If your transformation assumes metres (e.g., radial warps), ensure the input is in an appropriate projected CRS before calling this function.

Expected signature of `transform_fun`


transform_fun <- function(coords, ...) {  ## coords: n x 2 matrix (X, Y)
  ## return an n x 2 matrix with transformed (X, Y)}

Examples

library(sf)

# A simple coordinate transformer: scale and shift
scale_shift <- function(coords, sx = 1, sy = 1, dx = 0, dy = 0) {
 X <- coords[, 1] * sx + dx
 Y <- coords[, 2] * sy + dy
  cbind(X, Y)
}

# POINT example
pt <- st_sfc(st_point(c(0, 0)), crs = 3857)
st_transform_custom(pt, transform_fun = scale_shift,
                    args = list(sx = 2, sy = 2, dx = 1000, dy = -500))
#> Geometry set for 1 feature 
#> Geometry type: POINT
#> Dimension:     XY
#> Bounding box:  xmin: 1000 ymin: -500 xmax: 1000 ymax: -500
#> Projected CRS: WGS 84 / Pseudo-Mercator
#> POINT (1000 -500)

# LINESTRING example
ln <- st_sfc(st_linestring(rbind(c(0, 0), c(1, 0), c(1, 1))), crs = 3857)
st_transform_custom(ln, transform_fun = scale_shift,
                    args = list(sx = 10, sy = 10))
#> Geometry set for 1 feature 
#> Geometry type: LINESTRING
#> Dimension:     XY
#> Bounding box:  xmin: 0 ymin: 0 xmax: 10 ymax: 10
#> Projected CRS: WGS 84 / Pseudo-Mercator
#> LINESTRING (0 0, 10 0, 10 10)

# POLYGON example (unit square)
poly <- st_sfc(st_polygon(list(rbind(c(0,0), c(1,0), c(1,1),
                                     c(0,1), c(0,0)))), crs = 3857)
st_transform_custom(poly, transform_fun = scale_shift,
                    args = list(sx = 2, sy = 0.5, dx = 5))
#> Geometry set for 1 feature 
#> Geometry type: POLYGON
#> Dimension:     XY
#> Bounding box:  xmin: 0 ymin: 0 xmax: 1 ymax: 1
#> Projected CRS: WGS 84 / Pseudo-Mercator
#> POLYGON ((0 0, 1 0, 1 1, 0 1, 0 0))

# MULTIPOLYGON example (two disjoint squares)
mp <- st_sfc(st_multipolygon(list(
  list(rbind(c(0,0), c(1,0), c(1,1), c(0,1), c(0,0))),
  list(rbind(c(2,2), c(3,2), c(3,3), c(2,3), c(2,2)))
)), crs = 3857)
st_transform_custom(mp, transform_fun = scale_shift,
                    args = list(dx = 100, dy = 100))
#> Geometry set for 1 feature 
#> Geometry type: MULTIPOLYGON
#> Dimension:     XY
#> Bounding box:  xmin: 100 ymin: 100 xmax: 103 ymax: 103
#> Projected CRS: WGS 84 / Pseudo-Mercator
#> MULTIPOLYGON (((100 100, 101 100, 101 101, 100 ...

# In an sf data frame
sf_df <- st_sf(id = 1:2, geometry = st_sfc(
  st_point(c(10, 10)),
  st_linestring(rbind(c(0,0), c(2,0), c(2,2)))
), crs = 3857)

st_transform_custom(sf_df, transform_fun = scale_shift,
                    args = list(sx = 3, sy = 3))
#> Simple feature collection with 2 features and 1 field
#> Geometry type: GEOMETRY
#> Dimension:     XY
#> Bounding box:  xmin: 0 ymin: 0 xmax: 30 ymax: 30
#> Projected CRS: WGS 84 / Pseudo-Mercator
#>   id                   geometry
#> 1  1              POINT (30 30)
#> 2  2 LINESTRING (0 0, 6 0, 6 6)