coordinates

[dcooley]

General function to extract any dimensions, and any nesting of coordinates

---

TODO v0.1

• R-facing function
  • return data.frame with column names
• Vignette entry
• Tests

---

TODO v0.2

• fully templated
• can I get the dimensions of each individual geometry at the start, rather than in each iteration?
• optimise where all dimensions & nests are the same (to avoid extra dimensions() calls?)

[dcooley]

This is designed to be a general coordinate-flattener, where it can handle any level of nesting of matrices and lists.

l <- list(
  list(
    matrix(1:12, ncol = 2)
    , matrix(1:4, ncol = 2)
  )
  , 1:5
  , 1:2
  , matrix(1:9, ncol = 3)
)
gm_coordinates( l )
   id id1 id2 c1 c2 c3 c4 c5
1   0   1   1  1  7 NA NA NA
2   0   1   1  2  8 NA NA NA
3   0   1   1  3  9 NA NA NA
4   0   1   1  4 10 NA NA NA
5   0   1   1  5 11 NA NA NA
6   0   1   1  6 12 NA NA NA
7   0   2   2  1  3 NA NA NA
8   0   2   2  2  4 NA NA NA
9   1  NA  NA  1  2  3  4  5
10  2  NA  NA  1  2 NA NA NA
11  3  NA   1  1  4  7 NA NA
12  3  NA   1  2  5  8 NA NA
13  3  NA   1  3  6  9 NA NA

But, it's still slower because of this generalisation (I believe)

library(sf)
library(sfheaders)
# remotes::install_github("paleolimbot/wk")
library(wk)


nc_sf <- sf::read_sf(system.file("shape/nc.shp", package="sf"))
nc_sfc <- sf::st_geometry(nc_sf)
nc_wkb <- sf::st_as_binary(nc_sfc)


wk_wkb = wk::wkb_coords(rep(nc_wkb, 10))
sfheaders = sfheaders::sfc_to_df(rep(nc_sfc, 10))
sf = sf::st_coordinates(rep(nc_sfc, 10))

wb <- rep(nc_wkb, 10)
sfc <- rep(nc_sfc, 10)

microbenchmark::microbenchmark(
  wk_wkb = { df_wk <- wk::wkb_coords(wb) },
  sfheaders = { df_sf <- sfheaders::sfc_to_df(sfc) },
  geometries = { df_gm <- geometries:::rcpp_coordinates( sfc ) }
)

# Unit: milliseconds
#       expr      min       lq     mean   median        uq       max neval
#     wk_wkb 1.434450 1.526489 1.820761 1.575800  1.748121  9.416171   100
#  sfheaders 4.661102 5.189502 8.213676 5.586032 11.717638 56.477026   100
# geometries 4.045773 4.294310 5.387846 4.456686  4.746964 12.798332   100