Compare various ways of loading multiple files

loop_bench.R

# Create test data
lapply(1:9, function(x) {
  write.csv(
    nycflights13::flights,
    paste0("data", x, ".csv")
  )
})


# List all files to iterate through
all_data_files <- list.files(pattern = "\\.csv$")


# Define looping functions
for_loop <- function(data_files) {
  data_list <- vector(mode = "list", length = length(data_files))
  for (file in data_files) {
    data_list[[file]] <- read.csv(file)
  }
  dplyr::bind_rows(data_list)
}


lapply_loop <- function(data_files) {
  dplyr::bind_rows(
    lapply(data_files, read.csv)
  )
}


map_loop <- function(data_files) {
  dplyr::bind_rows(
    purrr::map(data_files, read.csv)
  )
}


map_df_loop <- function(data_files) {
  purrr::map_df(data_files, read.csv)
}


# Do the benchmark
result <- bench::mark(
  for_loop(all_data_files),
  lapply_loop(all_data_files),
  map_loop(all_data_files),
  map_df_loop(all_data_files)
)

# My result
tibble::tribble(
  ~expression,                   ~min,     ~median,
  "for_loop(all_data_files)",    "8.65s",  "8.65s",
  "lapply_loop(all_data_files)", "8.91s",  "8.91s",
  "map_loop(all_data_files)",    "8.84s",  "8.84s",
  "map_df_loop(all_data_files)", "8.96s",  "8.96s"
)