Convert identifying variable to unique hash

Convert a column of unique but restricted IDs into a set of new IDs using secure (SHA-2) hashing algorithm. Users have the option of saving a crosswalk between the old and new IDs in case observations need to reidentified at a later date.

deid_dua(
  df,
  id_col = NULL,
  new_id_name = "id",
  id_length = 64,
  existing_crosswalk = NULL,
  write_crosswalk = FALSE,
  crosswalk_filename = NULL
)

Arguments

df	Data frame
id_col	Column name with IDs to be replaced. By default it is NULL and uses the value set by the id_column argument in set_dua_level() function.
new_id_name	New hashed ID column name, which must be different from old name.
id_length	Length of new hashed ID; cannot be fewer than 12 characters (default is 64 characters).
existing_crosswalk	File name of existing crosswalk. If existing crosswalk is used, then new_id_name, id_length, id_length, and crosswalk_name will be determined by the already existing crosswalk. Arguments given for these values will be ignored.
write_crosswalk	Write crosswalk between old ID and new hash ID to console (unless crosswalk_name is given value).
crosswalk_filename	Name of crosswalk file with path; defaults to generic name with current date (YYYYMMDD) appended.

Examples

## --------------
## Setup
## --------------
## set DUA crosswalk
dua_cw <- system.file('extdata', 'dua_cw.csv', package = 'duawranglr')
set_dua_cw(dua_cw)
#> -- duawranglr note -------------------------------------------------------------
#> DUA crosswalk has been set!
## read in data
admin <- system.file('extdata', 'admin_data.csv', package = 'duawranglr')
df <- read_dua_file(admin)
## --------------

## show identified data
df
#> # A tibble: 9 x 10
#>   sid         sname    dob      gender raceeth tid   tname zip   mathscr readscr
#>   <chr>       <chr>    <chr>    <chr>  <chr>   <chr> <chr> <chr> <chr>   <chr>  
#> 1 000-00-0001 Schaefer 19900114 0      2       1     Smith 22906 515     496    
#> 2 000-00-0002 Hodges   19900225 0      1       1     Smith 22906 488     489    
#> 3 000-00-0003 Kirby    19900305 0      4       1     Smith 22906 522     498    
#> 4 000-00-0004 Estrada  19900419 0      3       1     Smith 22906 516     524    
#> 5 000-00-0005 Nielsen  19900530 1      2       1     Smith 22906 483     509    
#> 6 000-00-0006 Dean     19900621 1      1       2     Brown 22906 503     523    
#> 7 000-00-0007 Hickman  19900712 1      1       2     Brown 22906 539     509    
#> 8 000-00-0008 Bryant   19900826 0      2       2     Brown 22906 499     490    
#> 9 000-00-0009 Lynch    19900902 1      3       2     Brown 22906 499     493    

## deidentify
df <- deid_dua(df, id_col = 'sid', new_id_name = 'id', id_length = 12)

## show deidentified data
df
#> # A tibble: 9 x 10
#>   id           sname    dob     gender raceeth tid   tname zip   mathscr readscr
#>   <chr>        <chr>    <chr>   <chr>  <chr>   <chr> <chr> <chr> <chr>   <chr>  
#> 1 a14856cc5d00 Schaefer 199001… 0      2       1     Smith 22906 515     496    
#> 2 a141ce13114a Hodges   199002… 0      1       1     Smith 22906 488     489    
#> 3 de520f632a2c Kirby    199003… 0      4       1     Smith 22906 522     498    
#> 4 889d833f94ed Estrada  199004… 0      3       1     Smith 22906 516     524    
#> 5 2993f4bda3cd Nielsen  199005… 1      2       1     Smith 22906 483     509    
#> 6 86c8de9a8d63 Dean     199006… 1      1       2     Brown 22906 503     523    
#> 7 cdb300787c0b Hickman  199007… 1      1       2     Brown 22906 539     509    
#> 8 ef91ae029e71 Bryant   199008… 0      2       2     Brown 22906 499     490    
#> 9 0fb2736cec2c Lynch    199009… 1      3       2     Brown 22906 499     493    

if (FALSE) {
## save crosswalk between old and new ids for future
deid_dua(df, write_crosswalk = TRUE)

## use existing crosswalk (good for panel datasets that need link)
deid_dua(df, existing_crosswalk = './crosswalk/master_crosswalk.csv')
}