Loop that restarts for rate limit - r

Question

I am querying Google Maps Roads API which only accepts a maximum of 100 coordinate pairs per request. Some of my input linestrings from routes, however, contain more than 100 segments.
I've written a sample loop which sends a request to the API but it is limited to lines with under 100 segments with if (nrow(routes$mat) <= 100) {...} where routes$mat is the matrix of coordinate pairs.
for (i in 1:length(routes)) {
if (nrow(routes$mat) <= 100) {
mat <- paste(apply(mat, 1, paste, collapse=","), collapse="|")
a <- "https://roads.googleapis.com/v1/snapToRoads?path="
b <- mat
c <- "&interpolate=false&key=YOUR_API_KEY"
request <- paste(a,b,c, sep="")
con <- curl(request)
open(con)
out <- readLines(con)
mydf <- fromJSON(out)
close(con)
output <- cbind(mydf$snappedPoints$location$longitude, mydf$snappedPoints$location$latitude)
}
My objective is to send all of the lines to the API - regardless of their length, but this means sending them piecemeal.
How would one adjust this loop to send only 100 at a time if mat is longer than 100 and then concatenate the results into a single output ?
So for example, if nrow(mat) = 250, there would be 3 outputs, the first with 100, the second with 100, and the third with 50.

Answer 1

I always find doing this kind of loop a bit messy, but sometimes it's needed.
For this answer I'm using my googleway package which handles the API call for you. I'm also using the tram_route data that comes with it. There are 55 rows, so I'm iterating every 10, but with a larger data set you just increase the by = 10 value.
library(googleway)
set_key("roads_api_key", api = "roads")
n <- nrow(tram_route)
subsets <- c(seq(1, n, by = 10), n)
iters <- length(subsets) - 1
## set up a data.frame to store the results
df_result <- data.frame(latitude = numeric(n),
longitude = numeric(n))
for (i in 1:(length(subsets)-1) ) {
if (i == iters[length(iters)]) {
idx <- subsets[i]:subsets[i+1]
} else {
idx <- subsets[i]:(subsets[i+1] -1)
}
print(idx)
res <- google_snapToRoads(df_path = tram_route[idx, ],
lat = "shape_pt_lat",
lon = "shape_pt_lon")
df_result[idx, ] <- res$snappedPoints$location
}
head(df_result)
# latitude longitude
# 1 -37.81436 144.9386
# 2 -37.81330 144.9415
# 3 -37.81274 144.9429
# 4 -37.81268 144.9430
# 5 -37.81314 144.9439
# 6 -37.81351 144.9443
And proof if proof be needed
set_key("map_api_key")
df_result$colour <- "blue"
google_map() %>%
add_markers(tram_route, lat = "shape_pt_lat", lon = "shape_pt_lon") %>%
add_markers(df_result, colour = "colour")