Speed up raw csv reads and writes in R

src/data/download_dataset.R

@@ -3,6 +3,7 @@ source("packrat/init.R")
 library(RMySQL)
 library(stringr)
 library(dplyr)
+library(readr)
 
 participant <- snakemake@input[[1]]
 group <- snakemake@params[["group"]]
@@ -44,5 +45,5 @@ if("device_id" %in% available_columns){
 } else {
     print(paste0("Table ", table, "does not have a device_id column (Aware ID) to link its data to a participant"))
 }
-write.csv(sensor_data, sensor_file, row.names = FALSE)
+write_csv(sensor_data, sensor_file)
 dbDisconnect(stopDB)

src/data/readable_datetime.R

@@ -1,6 +1,7 @@
 source("packrat/init.R")
 
 library("tidyverse")
+library(readr)
 
 input <- read.csv(snakemake@input[[1]])
 sensor_output <- snakemake@output[[1]]
@@ -19,7 +20,7 @@ split_local_date_time <- function(data){
                                                 local_hour %in% 18:23 ~ "evening")))
 }
 if(!is.null(timezone_periods)){
-    timezones <- read.csv(timezone_periods)
+    timezones <- read_csv(timezone_periods)
     tz_starts <- timezones$start
     output <- input %>% 
                 mutate(timezone = findInterval(timestamp / 1000, tz_starts), # Set an interval ID based on timezones' start column
@@ -36,5 +37,5 @@ if(!is.null(timezone_periods)){
                 mutate(utc_date_time = as.POSIXct(timestamp/1000, origin="1970-01-01", tz="UTC"),
                         local_date_time = format(utc_date_time, tz = fixed_timezone, usetz = F))
     output <- split_local_date_time(output)
-    write.csv(output, sensor_output, row.names = FALSE)
+    write_csv(output, sensor_output)
 }