Deduplicate and disambiguate an attribute using both lag and lead values

Handles missing values and cases where the same date has conflicting attribute values by inferring from the closest neighbor (lag first, then lead).

Usage

dedup_value_panel(data, attr_col, id_col, date_col)

Arguments

data

A data frame.

attr_col

Attribute to disambiguate (unquoted).

id_col

Unique identifier column (unquoted).

date_col

Date column for ordering (unquoted).

Value

A deduplicated data frame with attribute filled from neighbors when missing or conflicting.

Examples

library(tibble)

df <- tibble(
  personnel_id = c(1, 1, 1, 2, 2, 2, 2),
  gender    = c(NA, "M", "F", "F", NA, "M", "F"),
  ref_date  = as.Date(c(
    "2023-01-01", "2023-01-02", "2023-01-02",
    "2023-01-01", "2023-01-02", "2023-01-03", "2023-01-03"
  ))
)

dedup_value_panel(df, gender, personnel_id, ref_date)
#> # A tibble: 5 × 3
#>   personnel_id ref_date   gender
#>          <dbl> <date>     <chr> 
#> 1            1 2023-01-01 NA    
#> 2            1 2023-01-02 F     
#> 3            2 2023-01-01 F     
#> 4            2 2023-01-02 NA    
#> 5            2 2023-01-03 F