For each row of df, picks the child's CDI score closest in age in each
language, form and measure, among those at most max_age_diff months away.
Arguments
- df
A table with
subject_idandage, e.g.pb$administrations- cdi
CDI scores, one per row, e.g. from
unnest_aux(pb$subjects, "cdi_responses"). Needssubject_id,age,language,instrument_typeandmeasure.- max_age_diff
The furthest a CDI score may be from the row's age, in months
- prefer_later
Of a score before and one after the row's age that are equally close, pick the later one instead of the earlier one
Value
df with the columns of cdi added, one row per row of df,
language, form and measure, in the order of df. The score's age is
cdi_age. Columns cdi shares with df, like dataset_id, appear once
and have to hold the same values in both for each subject. Rows without a
score close enough are left out.
Examples
if (FALSE) { # \dontrun{
pb <- peekbank("2026.1")
cdi <- unnest_aux(pb$subjects, "cdi_responses")
nearest_cdi(pb$administrations, cdi, max_age_diff = 3)
} # }