Skip to contents

Compute field-level diffs between two roller state tibbles

Source: R/roller-diff.R
diff_roller_state.Rd

Pure function: takes two flattened roller tibbles (as produced by flatten_roles() or parse_roles_bulk()) and returns a long-format changelog recording every field-level mutation. Detects three categories of change: role additions, role removals, and field-level modifications on continuing roles.

Usage

diff_roller_state(
  old_state,
  new_state,
  timestamp = format(Sys.time(), "%Y-%m-%dT%H:%M:%S"),
  update_id = NA_integer_
)

Arguments

old_state

Tibble. Previous roller state from read_state() or an earlier flatten_roles() call. NULL or zero-row tibble treats all current roles as additions.

new_state

Tibble. Current roller state from brreg_download() or flatten_roles().

timestamp

Character or POSIXct. Timestamp for changelog entries (typically the CDC event time or sync time).

update_id

Integer or character. Identifier for the sync batch, used as update_id in the changelog.

Value

A tibble matching the changelog schema: timestamp, org_nr, registry (always "roller"), change_type ("entry", "exit", "change"), field, value_from, value_to, update_id.

Composite key

Each role assignment is identified by (org_nr, role_group_code, role_code, holder_id) where holder_id is derived from person_id for person-held roles and entity:{entity_org_nr} for entity-held roles (auditors, accountants). Roles with neither are keyed as unknown:{row_index} within their respective state — these are rare and produce conservative add/remove pairs rather than false modifications.

NA handling

For additions, fields that are NA in the new state are excluded from the changelog (no value to report). For removals, fields that are NA in the old state are excluded. For modifications, a change from NA to a non-NA value (or vice versa) is recorded.

See also

brreg_sync() for automated sync with changelog persistence, brreg_roles() for fetching current roles, brreg_changes() for querying stored changelogs.

Other tidybrreg data management functions: brreg_annotation_summary(), brreg_annotations(), brreg_status(), brreg_sync(), brreg_sync_status(), read_changelog()

Examples

if (FALSE) { # interactive() && curl::has_internet()
# Detect board changes for a single company
old <- brreg_roles("923609016")
Sys.sleep(1)
new <- brreg_roles("923609016")
diff_roller_state(old, new)

# Bootstrap: NULL old state treats all roles as entries
roles <- brreg_roles("923609016")
diff_roller_state(NULL, roles)
}