read_csv saves geometry point data as character

[sj-io]

Hello everyone! I have a seemingly basic problem I can't figure out.

I have a csv file and one of the columns contains geometric point data. All the values in the column are formatted like:

POINT (-12.345 67.8910)
POINT (-12.543 67.0198)

I'm using read_csv to import the dataset, and it keeps assigning the column type as character. I eventually want to plot the data using sf, but st_read(), st_point(), & st_as_sf() functions require the col type to be a numeric matrix .

How do I correct the column type or is there something easy I'm missing?

[jlacko]

The values of your column seem to be in the well known text standard, and you can transform them to {sf} format using st_as_sf() and specifying the wkt and crs arguments.

So consider this example:

library(sf)
library(dplyr)


raw <- data.frame(label = c("point A",
                            "point B"),
                  geometry = c("POINT (-12.345 67.8910)",
                               "POINT (-12.543 67.0198)"))


clean <- raw %>% 
  st_as_sf(wkt = "geometry", # name of the column containg the WKT data
           crs = 4326) # coordinate reference system to make sense of the values

mapview::mapview(clean)

Note that I am making an assumption that your coordinates are lat/long in degrees; in the age of GPS this is arguably the most frequent case, but it still is an assumption on my side.

[sj-io]

This worked! Thank you!