convert csv to shp in r where geometry is in single column-CodePudding

I have a csv file that contains the point geometry information in a single column. Is there a straight forward way to convert from csv to a spatial data file in r given the format of the geometry column (I could do this in QGIS or could splice the column into x and y, but I'm curious if there is a better way to do it).

Here is an example of what the data look like:

name <- c("A", "B", "C")
geom <- c("POINT (45.095914704767 -93.266719775361)",
          "POINT (45.095220489232 -93.254896591796)",
          "POINT (45.079643666 -93.257941333)")
dat <- data.frame(name, geom)
dat

CodePudding user response：

If your geometry column is actually formatted as a string, you could use dplyr to strip away the excess text and then use the sf package to convert the coordinates into a point column:

library(magrittr)
dat %>%
  dplyr::mutate(
    # replace text and parenthesis
    geom = stringr::str_replace(geom, 'POINT \\(', ''),
    geom = stringr::str_replace(geom, '\\)', '')
  ) %>%
  # separate into lat and lon columns
  tidyr::separate(geom, into=c('lon', 'lat'), sep=' ') %>%
  # convert to sf point object 
  # (assuming this is in WGS84, but you can specify any CRS here)
  sf::st_as_sf(coords = c('lat', 'lon'), crs=4326)

If you are able to save your .csv as a .geojson or a .shp file, you can just read it into R with the sf::read_sf('path/to/your/data.shp') function.

CodePudding user response：

With base regex (and sf):

library(sf)

dat$y <- gsub(pattern = ".*\\((-?[0-9.] ).*", replacement= "\\1", dat$geom)
dat$x <- gsub(pattern = ".*\\s(-?[0-9.] ).*", replacement= "\\1", dat$geom)
dat_sf <- st_as_sf(dat, coords = c("y","x"))
st_write(dat_sf, "dat.shp")

Created on 2021-10-05 by the reprex package (v2.0.1)