In Epiconcept-Paris/STRAP-epiuf: Epiconcept Team Utility Functions for STRAP Project

knitr::opts_chunk$set(echo = TRUE)
library(epiuf)
library(kableExtra)

Epifunction : getDelayIf

This document show examples of epiFunction use

Tests

Test when the data is not a dataframe:

# Create a random table with two date variables
nbRecords <- 50
rand_tbl <- tibble::as_tibble(data.frame(
  date1 = sample(seq(as.Date('2000-01-01'), as.Date('2022-01-01'), by = "day"), nbRecords),
  date2 = sample(seq(as.Date('2000-01-01'), as.Date('2022-01-01'), by = "day"), nbRecords)
))


rand_tbl$delayDate1Date2 <- getDelayIf(rand_tbl, date1, date2)

Test when used within other functions:

# Create 2 dummy data sets with dates and ids----
df <- data.frame(date1 = c("2023-01-01", "2023-01-02", "2023-01-03", "2023-01-04")
                 , date2 = c("2023-02-01", "2023-02-02", "2023-04-03", "2023-03-04")
                 , id = c("a", "b", "c", "d"))

df2 <- data.frame(date1 = c("2023-01-01", "2023-01-02", "2023-01-03")
                  , date2 = c("2023-02-01", "2023-02-02", "2023-04-03")
                  , id = c("a", "b", "c"))

df$date1 <- as.Date(df$date1)
df$date2 <- as.Date(df$date2)
df2$date1 <- as.Date(df2$date1)
df2$date2 <- as.Date(df2$date2)


# Situation 1: use within a with() function----
# This works fine (no with())
df$test <- ifelse(getDelayIf(df, date1, date2)>40, 1, 0) 

#This also works fine (not using getDelayIf so that can use with())
df$test2 <- with(df, ifelse(!is.na(date1) & !is.na(date2) & date2-date1>40, 1, 0)) 

# This does not work as data must be specified within getDelayIf 
# It would be useful if the getDelayIf could be designed such that it can be used in this setting.
#df$test <- with(df, ifelse(getDelayIf(FirstDateName = date1, SecondDateName = date2)>40, 1, 0)) 


# Situation 2: using within subset()----
# This works, not using getDelayIf -> retrieve 2 rows as expected
df_test <- subset(df,!is.na(date1) & !is.na(date2) & date2-date1>40 )
# This also works, though it seems ineficient to specify df twice
df_test2 <- subset(df,getDelayIf(df, date1, date2)>40)

# This does not work, as must specify df in getDelayIf.
#df_test2 <- subset(df,getDelayIf(date1, date2)>40)

# Situation 2: use within printDataCheck when 2 data sets are being assessed----
# printDataCheck parses a condition through subset() for each of data and data_old in turn.
# As the same condition is used for both data sets, no specific dataset can be listed in the condition.

# Create example data checks, one using getDelayIf, the other not
check <- list("!is.na(date1) & !is.na(date2) & getDelayIf(df, date1, date2)>40" = "date difference greater than 40 days"
              ,"!is.na(date1) & !is.na(date2) & date2-date1>40" = "test date difference greater than 40 days")

# Run in printDataCheck:
#This produces the expected output, but results in a warning
#as have to specify one data set in getDelayIf
#BUT the printDataCheck is running the condition on both df AND df2

printDataCheck(data=df, data_old=df2, parameters = check, varname = "id", header = "Testing getDelayIf")

#Warning message:
#In !is.na(date1) & !is.na(date2) & getDelayIf(df, date1, date2) >  :
#  longer object length is not a multiple of shorter object length

# Conclusion: 
# It would be useful if the getDelayIf could be designed such that it can be used in these settings WITHOUT the need to specify the data.

Epiconcept-Paris/STRAP-epiuf documentation built on Aug. 5, 2024, 3:41 a.m.