Separate clean cols

[zkamvar]

This is in accordance with @dirkschumacher's suggestion in #1 Also, I found the comment() function, which seems really useful for this task :)

What it does:

• clean_* functions no longer overwrite data explicitly. They now create *_clean columns that also contain a "<linelist>clean" comment to prevent other clean_* functions from overwriting them.
• when no changes occur, a *_clean column is not created and no label is applied
• reduces the number of operations by only acting on the unclean columns

@thibautjombart, @dirkschumacher, what do you think?

[zkamvar]

To clarify, this is what the process looks like:

library("linelist")

onsets <- as.Date("2018-01-01") + sample(1:10, 20, replace = TRUE)
discharge <- format(as.Date(onsets) + 10, "%d/%m/%Y")
genders <- c("male", "female", "FEMALE", "Male", "Female", "MALE")
gender <- sample(genders, 20, replace = TRUE)
case_types <- c("confirmed", "probable", "suspected", "not a case",
                "Confirmed", "PROBABLE", "suspected  ", "Not.a.Case")
messy_dates <- sample(
                 c("01-12-2001", "male", "female", "2018-10-18", "2018_10_17",
                   "2018 10 19", "// 24//12//1989", NA, "that's 24/12/1989!"),
                 20, replace = TRUE)
case <- factor(sample(case_types, 20, replace = TRUE))
toy_data <- data.frame("Date of Onset." = onsets,
                       "DisCharge.." = discharge,
                       "GENDER_ " = gender,
                       "Épi.Case_définition" = case,
                       "messy/dates" = messy_dates)
head(toy_data)
#>   Date.of.Onset. DisCharge.. GENDER_. Épi.Case_définition
#> 1     2018-01-08  18/01/2018   FEMALE            PROBABLE
#> 2     2018-01-06  16/01/2018   female          Not.a.Case
#> 3     2018-01-04  14/01/2018     male           Confirmed
#> 4     2018-01-04  14/01/2018   female           Confirmed
#> 5     2018-01-08  18/01/2018   FEMALE           confirmed
#> 6     2018-01-05  15/01/2018   female         suspected  
#>          messy.dates
#> 1         2018-10-18
#> 2    // 24//12//1989
#> 3         2018 10 19
#> 4 that's 24/12/1989!
#> 5 that's 24/12/1989!
#> 6             female
head(cd <- clean_data(toy_data))
#>   date_of_onset  discharge gender epi_case_definition        messy_dates
#> 1    2018-01-08 18/01/2018 FEMALE            PROBABLE         2018-10-18
#> 2    2018-01-06 16/01/2018 female          Not.a.Case    // 24//12//1989
#> 3    2018-01-04 14/01/2018   male           Confirmed         2018 10 19
#> 4    2018-01-04 14/01/2018 female           Confirmed that's 24/12/1989!
#> 5    2018-01-08 18/01/2018 FEMALE           confirmed that's 24/12/1989!
#> 6    2018-01-05 15/01/2018 female         suspected               female
#>   discharge_clean messy_dates_clean gender_clean epi_case_definition_clean
#> 1      2018-01-18        2018-10-18       female                  probable
#> 2      2018-01-16        1989-12-24       female                not_a_case
#> 3      2018-01-14        2018-10-19         male                 confirmed
#> 4      2018-01-14        1989-12-24       female                 confirmed
#> 5      2018-01-18        1989-12-24       female                 confirmed
#> 6      2018-01-15              <NA>       female                 suspected
lapply(cd, comment)
#> $date_of_onset
#> NULL
#> 
#> $discharge
#> NULL
#> 
#> $gender
#> NULL
#> 
#> $epi_case_definition
#> NULL
#> 
#> $messy_dates
#> NULL
#> 
#> $discharge_clean
#> [1] "<linelist>clean"
#> 
#> $messy_dates_clean
#> [1] "<linelist>clean"
#> 
#> $gender_clean
#> [1] "<linelist>clean"
#> 
#> $epi_case_definition_clean
#> [1] "<linelist>clean"

^{Created on 2019-01-18 by the reprex package (v0.2.1)}

[zkamvar]

I'm going to close this PR since it is woefully out of date and will not be merged