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

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

Description:

validDate function aims to convert an character date into the Universal Date Format (UDF) in R (YYYY-mm-dd)

validDate can be use with various date format, the function will guess the format by analysing the first values of the character date

Workflow:

This function workflow is divided into 5 phases:

• PHASE 1: Prints warnings if any date inside the date vector is too short or if > 10% of the date vector values have too short length. In these cases: validDate converts them to NA, as it's impossible to guess date format
  
• Example 1: datevar <- c(...,"3/3/3",...)
  

• Example 2: datevar <- c("3/3/3","555","03031")
  

• PHASE 2: Guess if:
  a) There are separators in the date vector -> function selects the most frequent separator
  b) Or, it's compound of only digits -> parameter digits = TRUE (kind of flag to know there are only digits within datevar)
  

• PHASE 3: Save different parts of the date depending on:
  a) There are separators in datevar -> create part where part[,1], part[,2] and part[,3] contains day,month,year (not knowing yet which one is)
  b) It's compound of only digits -> as there is no separator, just guess if the year is at the beginning or at the end and save 3 parts too
  

• PHASE 4: Guess which part is day,month,year depending on the max, min and range values
  

• PHASE 5: Assemble datevar to call function as.Date(x = datevar, format = the format guessed)
  

Testing the function

# Example 1. The 3rd date is converted to NA because looking at the overall vector the number 15 doesn't corresponds to any month number.
dates1 <-  c("2022/01/15","2022/02/10","1950/01/31","2022/15/15",NA)
validDate(dates1)
# Example 2. A only-digit vector.
dates2 <- c("20221120","20210615","20210303","",NA)
validDate(dates2)
# Example 3. Using the same vector as dates1 but with "." as the separator.
dates3 <- c("2022.01.15","2022.02.10","1950.01.31","2022.15.15",NA)
validDate(dates3)
# Example 4. Using dates1 vector but with a blank space a the separator.
dates4 <- c("2022 01 15","2022 02 10","1950 01 31","2022 15 15",NA)
validDate(dates4)
# Example 5. A vector in which is not possible to guess which position carries the year, the month and the day returns NA.
dates5 <- c("1/2/3","1/3/22","1/2/3","1/2/3","1/2/3","1/2/3","1/2/3")
validDate(dates5)
# Example 6. A unique value in which is not possible to guess the year, month and date will return NA.
validDate(c("1/1/2"))
# Example 7. A unique value in which the year, month and date are predictable will return the vector converted into R Date format.
validDate("30/1/2022")
# Example 8. A vector in which the second date doesn't exists (30 february) gives NA for that value. 
validDate(c("2010/3/20","2010/2/30"))
# Example 9. A vector dated on the future.
validDate(c("21.05.2030","01.11.2050","15.07.2024"))
# Example 10. An ambiguous vector in terms of year. If we are sure that all the dates are in the past, we can use the parameter dropFuture = TRUE.
validDate(c("21.05.30","01.11.50","31.11.25"),dropFuture = T)
# Example 11. A vector with mixed formats. The function will convert into NA the formats that are not represented by the majority in the vector.
dates11 <- c("17/10/2021","21.10.2020","07.03.2021")
validDate(dates11)

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