Open kadyb opened 1 year ago
I wonder if it would be resonable to only return X and Y coordinates for points when converting geometry to data frame like sf::st_coordinates() does?
sf::st_coordinates()
library("sf") library("sfheaders") n = 100000 df = data.frame(x = rnorm(n), y = rnorm(n)) pts = st_as_sfc(st_as_sf(df, coords = c("x", "y"))) x = sf::st_coordinates(pts) str(x) #> num [1:100000, 1:2] -0.4177 -0.0679 0.6598 0.756 0.1694 ... #> - attr(*, "dimnames")=List of 2 #> ..$ : NULL #> ..$ : chr [1:2] "X" "Y" format(object.size(x), units = "auto") #> [1] "1.5 Mb" y = sfheaders::sfc_to_df(pts) str(y) #> 'data.frame': 100000 obs. of 4 variables: #> $ sfg_id : int 1 2 3 4 5 6 7 8 9 10 ... #> $ point_id: int 1 2 3 4 5 6 7 8 9 10 ... #> $ x : num -0.4177 -0.0679 0.6598 0.756 0.1694 ... #> $ y : num -0.71246 -0.00779 -2.32569 -0.14325 2.83237 ... format(object.size(y), units = "auto") #> [1] "2.3 Mb" t = bench::mark( check = FALSE, sf = st_coordinates(pts), sfheaders = sfc_to_df(pts), ) t[, 1:5] #> expression min median `itr/sec` mem_alloc #> 1 sf 1.72ms 1.78ms 532. 3.05MB #> 2 sfheaders 28.15ms 29.07ms 34.1 4.2MB
I wonder if it would be resonable to only return X and Y coordinates for points when converting geometry to data frame like
sf::st_coordinates()does?