Using st_read to import large geojson in iterations R

问题

I want to import a very large geojson into a simple features object using st_read in R, however the hardware demands seem to be large when converting from geojson to sf. for example, importing the microsoft data for building footprints for Ohio (https://github.com/Microsoft/USBuildingFootprints) which is a 1.2 GB geojson eats up over 32 GB of RAM when converting. Is their a method for iterating through rows of a geojson in a function so I can import parts of the whole file without eating up all that RAM, similar to skip rows in read.csv?

回答1:

Using library(geojsonsf) seems to work without issue on my Mac with 16gb RAM

library(geojsonsf)
library(sf)

sf <- geojsonsf::geojson_sf("~/Downloads/Ohio.geojson")

sf
# Simple feature collection with 5449419 features and 0 fields
# geometry type:  POLYGON
# dimension:      XY
# bbox:           xmin: -84.82027 ymin: 38.40334 xmax: -80.51887 ymax: 41.97041
# epsg (SRID):    4326
# proj4string:    +proj=longlat +datum=WGS84 +no_defs
# First 10 features:
#   geometry
# 1  POLYGON ((-84.81222 39.9087...
# 2  POLYGON ((-84.80084 39.8882...
# 3  POLYGON ((-84.78565 39.8811...
# 4  POLYGON ((-84.7373 39.9014,...
# 5  POLYGON ((-84.73916 39.8980...
# 6  POLYGON ((-84.80422 39.8646...
# 7  POLYGON ((-84.80025 39.8592...
# 8  POLYGON ((-84.79336 39.8593...
# 9  POLYGON ((-84.79268 39.8604...
# 10 POLYGON ((-84.80194 39.8639...

来源：https://stackoverflow.com/questions/52899355/using-st-read-to-import-large-geojson-in-iterations-r