Add the 0.12 species files to a built obis-h3 release — obis_h3_amend

Amends a release written by obis_h3_export_parquet() 0.11 (or later) in place, from its own files and the store's eov table, without re-running the export:

Usage

obis_h3_amend_species(
  con,
  release_dir,
  from_res = 3L,
  res_flat = c(1L, 2L),
  split = c(`3` = 1L),
  eov = "eov",
  row_group_rows = 10000L,
  split_row_group_rows = 2048L,
  manifest = TRUE,
  compression = "zstd"
)

Arguments

con: a DBI connection to DuckDB with the h3 extension loaded (the store, for its eov table).
release_dir: the release directory.
from_res: the resolution of the existing flat file (default 3).
res_flat: the coarser resolutions to roll up to (default 1 and 2).
split: resolution = parent resolution of the split (default c("3" = 1L)).
eov: the EOV membership table, or NULL for no eov column.
row_group_rows: rows per row group of the roll-ups, the lookup and the lineage file (default 10000).
split_row_group_rows: rows per row group of the split (default 2048).
manifest: also update files.parquet (default TRUE).
compression: Parquet codec (default "zstd").

Value

invisibly, a data frame of the files written (path, rows, bytes), with the attribute eov (aphiaids per EOV).

Details

species/species.parquet rewritten with the column eov (obis_h3_species_eov_sql(); same rows, same order) and species/lineage.parquet beside it (obis_h3_write_species_lookup());
species_flat/res=<r>.parquet for res_flat and the split species_flat/res=<s>/p=<parent>.parquet (obis_h3_write_species_flat()), from species_flat/res=<from_res>.parquet, which is not rewritten;
release.json: layers.species_flat (resolutions, the split's template_parent, parent_res, by_res, split), layers.species (lookup_columns with eov, lineage, eov, lookup_bytes) and, with manifest = TRUE, files.parquet (the lookup's new size, the new files added; every other row kept).

The collections are not written here: run obis_h3_stac_collection() and obis_h3_stac_stable() after.