Add harmonized coded occupation and industry columns — add_coded

Coded occupation and industry appear in the MCOD files under two different, **non-comparable** coding schemes at different byte positions and column names. This helper hides those details: given an imported MCOD data frame and its data year, it appends standardized columns (`occ_coded`, `ind_coded`, their recodes, the scheme, and an availability flag) so downstream code does not need to know which era or tier the data came from.

Usage

add_coded_occupation(df, year)

Arguments

df: a data frame imported via [import_mcod_fwf()] (or `.import_restricted_data()`)
year: the data year of `df` (integer)

Value

`df` with added columns: `occ_scheme` (character, the coding scheme or `NA`), `occ_coded`, `ind_coded`, `occ_recode`, `ind_recode`, and `occ_available` (logical; `TRUE` when the scheme applies and `occ_coded` has any non-missing value)

Details

Availability matrix (coded occupation/industry):

Years	Scheme	Notes
1982-1999	`3digit_census`	1980-Census basis 1982-1992, 1990-Census 1993-1999; source columns `occup` (@88-90) and `industry` (@85-87); state-dependent coverage
2000-2019	(none)	coded occupation/industry not collected
2020+	`4digit_niosh`	NCHS+NIOSH 4-digit codes; source columns `occupation`/`occupationr` (@806-811) and `industry`/`industryr` (@812-817)

The 3-digit (1982-1999) and 4-digit (2020+) codes are **not comparable** – do not chain a series across the gap. Tier difference: the 4-digit codes reach the public file in data year 2020 but the restricted file only in 2021, so `occ_coded` is all-`NA` for restricted 2020 (use the public file for 2020 occupation); from 2021 the public and restricted codes are identical.

Examples

if (FALSE) { # \dontrun{
df <- import_mcod_fwf("mort2023us.dat", 2023, tier = "public")
df <- add_coded_occupation(df, 2023)
table(df$occ_scheme)          # "4digit_niosh"
} # }