# last updated 2022-05-03

# This is example R code to link event and condition records from an IPUMS MEPS
# hierarchical extract

# Load the ipumsr package. If you do not ipumsr installed already you will need
# to use the command
#install.packages("ipumsr")
library(ipumsr)

# Read your IPUMS MEPS extract into R using the ipumsr package - UPDATE HERE
setwd("C:\user\directory") # file where your extract is saved
ddi <- ipumsr::read_ipums_ddi("meps_00001.xml") # your extract name
data <- ipumsr::read_ipums_micro(ddi)

# Split the hierarchical extract into different objects for each record type
for (i in unique(data$RECTYPE)) {
  data_rectype <- data[data$RECTYPE == i, ]
  cols_keep <-
    vapply(data_rectype, FUN.VALUE = logical(1),
           FUN = function (x) sum(!is.na(x)) != 0)
  cols_keep <- names(cols_keep)[cols_keep]
  data_rectype <- data_rectype[, cols_keep]
  assign(paste0("data_", i), data_rectype)
  rm(i, data_rectype, cols_keep)
}

# Link the condition and event records together using the condition-event link record
link_condition_events <-
  dplyr::inner_join(x = data_X,
  		    y = data_J,
  		    by = c("MEPSCONDID" = "MEPSCONDIDJ", "YEARX" = "YEARJ"))

link_condition_events <-
  dplyr::inner_join(x = link_condition_events,
                    y = data_E,
                    by = c("MEPSEVNTIDJ" = "MEPSEVNTID", "YEARX" = "YEARE"))

# The resulting link_condition_events data frame will have a record for each
# event-condition pair and will have filtered out all conditions that don't
# link to an event and all events that don't link to a condition