Build a PRIO-GRID hive-partitioned dataset — build_pg

Memory-safe alternative to read_pg_timevarying() that writes the hive partitioned parquet dataset and CSV bundle without loading the full table. Useful for manual rebuild and re-upload workflows.

Usage

build_pg_dataset(
  config = NULL,
  version = NULL,
  type = "05deg_yearly",
  spatial_hash = NULL,
  temporal_hash = NULL,
  overwrite = FALSE
)

Arguments

config: A pg_config object for custom data, or NULL (default) for the official release.
version: Character string specifying PRIOGRID version (release mode only).
type: Character string specifying release type. Default: "05deg_yearly".
spatial_hash: 6-character MD5 hash of spatial options (advanced use).
temporal_hash: 6-character MD5 hash of temporal options (advanced use).
overwrite: Logical. If TRUE, rebuilds even if output already exists.

Value

Invisibly returns the base output path.