Conversion of epoch time (string) to human readable date in R script

I am retrieving data from an ArcGIS Rest service using a web request in an R script. One of the attribute columns does contain a date value. These dates are returned as epoch values (character strings). I try to convert these epoch strings to human readable dates, but until now to no avail...

See my reproducible script below. The column containing the date is called WVK_BEGDAT. The content of this column should be converted to human readable dates.

I have tried several suggestions found via Google...

library(httr)
library(sf)
library(lubridate)

url <- parse_url("https://services.arcgis.com/nSZVuSZjHpEZZbRo/arcgis/rest/services")
url$path <- paste(url$path, "NWB_Wegvakken/FeatureServer/0/query", sep = "/")
url$query <- list(where = "WEGBEHSRT = 'R' AND WEGNUMMER = '015'",
                  outFields = "*",
                  returnGeometry = "true",
                  f = "json")
request <- build_url(url)
request
NWB <- st_read(request, stringsAsFactors = FALSE)
plot(st_geometry(NWB))

NWB$WVK_BEGDAT2 <- as_date(NWB$WVK_BEGDAT, format="%d-%m-%Y", tz = "CET")

The closest I get is using the as_date() function from the lubridate package. This function at least does not return an error, bu fills the column with NAs.

Any suggestions? Thanks in advance.

Solution

as.POSIXct should do the trick:

as.POSIXct(as.numeric(NWB$WVK_BEGDAT)/1000, origin = "1970-01-01")

You have to divide by 1000, because your epoch times seem to be in milliseconds.