在R中将SF包中的函数应用于每一行

Question

I'm working with the sf package in R, and for that purpose I'm trying to use one of their functions to create lines. The goal is to apply their functions to each row (columns 4:3 and 6:5) from the example data frame below.

df <- read.table(text = " from to from_lat from_lon to_lat to_lon travel_time
1  8015345 8849023 50.77083 6.105277 50.71896 6.041269    7.000000
2  8200100 8200101 49.60000 6.133333 49.63390 6.136765    8.000000
3  8200100 8200110 49.60000 6.133333 49.74889 6.106111   16.000000
4  8200100 8200510 49.60000 6.133333 49.61111 6.050000    4.857143
5  8200100 8200940 49.60000 6.133333 49.55129 5.845025   28.236842
6  8200100 8866001 49.60000 6.133333 49.68053 5.809972   37.000000
7  8200100 8869054 49.60000 6.133333 49.64396 5.904150   14.000000
8  8200101 8200100 49.63390 6.136765 49.60000 6.133333    7.000000
9  8200101 8200110 49.63390 6.136765 49.74889 6.106111   11.000000
10 8200110 8200100 49.74889 6.106111 49.60000 6.133333   17.074074", header = TRUE)

I know how to do this for one row:

library(sf)
library(dplyr)

x = matrix(as.numeric(c(df[1, c(4, 3)],
             df[1, c(6, 5)])), ncol = 2, byrow = TRUE)
class(x)
typeof(x)
l1 = st_linestring(x = x)
lsf = l1 %>% 
  st_sfc() %>% 
  st_sf(crs = 4326)
plot(lsf) #just to confirm that it is a line

But what I really need is to do it for every row. I tried to use a for loop but for some reason it messes up with the sf package classes. So I would assume that the solution would involve apply() but I'm not sure how to.

Answer 1

If we need to do this on each row, then we can use pmap

library(purrr)
library(dplyr)
df%>%
  select(4, 6, 3, 5) %>% 
  pmap(~ c(...) %>% 
            matrix(., ncol = 2) %>% 
            st_linestring %>%
            st_sfc %>%
            st_sf(crc = 4326))

Answer 2

You could use dplyr::rowwise() to do the row grouping.

df %>% rowwise() %>%
  mutate(line_sf = list(matrix(c(from_lon, to_lon, from_lat, to_lat), ncol = 2) %>%
           st_linestring()) ) %>%
   with(st_sfc(line_sf, crs = 4326)) %>%
   plot()

I re-arranged the last line (before plot ) to collapse the 10 line observations into a single geometry set for plotting here, but you could just leave them as individuals.

Answer 3

A solution that doesn't require dplyr:

lsf <- mapply(function(a, b, c, d) {
  list(matrix(c(a, b, c, d), ncol = 2) %>%
    st_linestring())
  }, df$from_lon, df$to_lon, df$from_lat, df$to_lat) %>%
  st_sfc(crs = 4326)

plot(lsf)