Skip to contents

Reads each layer of reference_gpkg once, groups rows by vpuid, and writes one ref_vpu{id}.gpkg per VPU into <out_dir>/vpu_{id}/work/. When these files exist, refactor Step 0 is an instant skip: eliminating the per-VPU random read against the large source file.

Usage

presplit_reference(
  reference_gpkg,
  out_dir,
  vpuids = NULL,
  layers = c("flowpaths", "divides", "hydrolocations", "pois"),
  overwrite = FALSE
)

Arguments

reference_gpkg

Path to the multi-VPU reference GeoPackage.

out_dir

Parent output directory (same as run_pipeline).

vpuids

Character vector of VPU IDs to split out. NULL (default) splits all VPUs found in the flowpaths layer.

layers

Layers to extract. Default matches refactor Step 0.

overwrite

If TRUE, overwrite existing per-VPU slices. Default FALSE.

Value

Invisibly, a named character vector vpuid -> path for every file written.

Examples

if (FALSE) { # \dontrun{
presplit_reference(reference_gpkg = "reference_conus.gpkg",
                   out_dir        = "out",
                   vpuids         = c("01", "02"))
} # }