Import data pull

# Load in packages that will be used throughout the process
library(timeDate)
library(gdata)
library(plyr)
library(zoo)

# Set working directory to the one where you want to keep all generated output
setwd("F:/Dropbox (Dimagi)/Output")

# Bulk import raw data from # project domains in list
file_dir <- "F:/Dropbox (Dimagi)/Input" 
filelist <- dir(file_dir, pattern = ".csv", recursive = TRUE, all.files = TRUE, full.names = TRUE) 
f <- lapply(filelist, read.csv, header = TRUE)

# Subset raw data 
subset_vars <- c("domain", "caseid", "flwid", "starttime", "endtime")
for (i in seq_along(f)) {
  names(f)[i] <- paste(as.character(f[[i]]$domain[1]))
  f[[i]] <- f[[i]][subset_vars] # Think this step could be skipped based on what we get and what type of analysis is requested
  f[[i]]$domain.index <- rep(i, nrow(f[[i]]))  # Add domain index to each data frame for further index use when they get merged into one big file
}

f_merged <- do.call("rbind", f)