Apply using characters

Question

I'm working with sf objects and I have a question about applying user-defined functions to data frames. This is a silly example, but it is similar to what I'm trying to do with a more complicated problem. I have a data frame called names that has two columns, each with the name of a county. I also read in the North Carolina data included with the sf package.

require(sf)
name_1 <- c('Ashe','Alleghany','Surry')
name_2 <- c('Currituck','Northampton','Hertford')
names <- data.frame(cbind(name_1,name_2))
nc <- st_read(system.file("shape/nc.shp", package="sf"))

What I'm trying to do is create a function that goes down each row of my data frame names, takes the name out of the name_1 column and the name out of the name_2 column, and sees if they intersect using the geometric data in nc. I have:

check_intersection <- function(x){
  return(st_intersects(nc[nc$NAME== x$name_1,],nc[nc$NAME==x$name_2,],sparse = FALSE))
}
apply(names,1,check_intersection)

But this yields an error of Error in x$name_1 : $ operator is invalid for atomic vectors. How do I tell the function to get the character values out of the appropriate columns, for each row in names?

Could you not subset your sf object by each vector and then run an st_intersects or st_intersection on the two objects? Not understanding why you need them in the same df — elmuertefurioso, Sep 20 '17 at 17:17
You can't use `$` operator with vectors. Replace x$name_1 and the others with something like x[,1] and x$name_2 with x[,2] — csgroen, Sep 20 '17 at 17:19

score 1 · Accepted Answer · answered Sep 26 '17 at 20:00

1

Try

apply(names, 1, 
  function(x) 
    st_intersects(nc[nc$NAME == x[1],], nc[nc$NAME == x[2],], 
       sparse = FALSE)
)

In the anonymous function, x is a vector, not a data.frame, hence $ does not work.

answered Sep 26 '17 at 20:00

Edzer Pebesma

3,474
14
23

Apply using characters

1 Answers1