Write the species lookup with its EOV, and the split lineage file

Writes dir/species.parquet (the rows of lookup, ordered by aphiaid, with the column eov of obis_h3_species_eov_sql() after the lookup's own columns) and dir/lineage.parquet (aphiaid, name_key, lineage only, the same rows), both in row groups of row_group_rows, so a WoRMS subtree reads the lineage without the names and a list of species reads the names without the lineage. Without an eov table (eov = NULL) the column is left out. The rows are lookup's, one per aphiaid, ordered by aphiaid (the order obis_h3_export_parquet() writes); each file is written to a temporary name and renamed, so lookup may read the species.parquet it replaces.

Usage

obis_h3_write_species_lookup(
  con,
  lookup,
  dir,
  eov = "eov",
  row_group_rows = 10000L,
  compression = "zstd"
)

Arguments

con

a DBI connection to DuckDB.

lookup

a table name or a table function (read_parquet('...')) with the lookup columns of obis_h3_species_lookup_sql().

dir

output folder (the release's species/).

eov

the EOV membership table, or NULL (default "eov").

row_group_rows

rows per row group (default 10000).

compression

Parquet codec (default "zstd").

Value

a data frame, one row per file: path (relative to the release), rows, bytes, with the attribute eov (the number of aphiaids per EOV, NA for none).

Details

No IUCN Red List column is written: the store holds no Red List field (its occ_h3 and taxon tables carry none).