Identifiers are never inferred from matching bytes or nearby timestamps. Duplicate content retains separate identities and is flagged for review. A recording must be unique on the supplied keys; multiple assets require an explicit one-to-many relationship. Missing keys are always rejected.
Usage
pr_link_assets(
recordings,
assets,
by,
relationship = c("one-to-one", "one-to-many"),
allow_unmatched = FALSE
)Arguments
- recordings
Data frame with recording_id and the acquisition keys.
- assets
Data frame with asset_id and the same keys. Optional sha256 values are used only to flag duplicate content, never to assign identity.
- by
Non-empty character vector of shared key column names.
- relationship
Either one-to-one (default) or one-to-many.
- allow_unmatched
Keep unmatched assets with a status flag.
Value
One row per asset, with recording_id, asset_id, acquisition keys, link_status and duplicate_content. Unused recording IDs are an attribute.
Examples
recordings <- data.frame(
recording_id = c("R1", "R2"),
subject = c("S01", "S02"),
session = c(1, 1)
)
assets <- data.frame(
asset_id = c("A1", "A2"),
subject = c("S02", "S01"),
session = c(1, 1)
)
pr_link_assets(recordings, assets, by = c("subject", "session"))
#> # A tibble: 2 × 6
#> asset_id subject session recording_id link_status duplicate_content
#> <chr> <chr> <dbl> <chr> <chr> <lgl>
#> 1 A1 S02 1 R2 matched_explicit_keys FALSE
#> 2 A2 S01 1 R1 matched_explicit_keys FALSE