| Title: | Cloud-Optimized ZIP - Open a ZIP Like a Table |
|---|---|
| Description: | R bindings to libcozip, the reference library for the Cloud-Optimized ZIP (cozip) format. |
| Authors: | Cesar Aybar [aut, cre], Roy Yali [aut], Julio Contreras [aut] |
| Maintainer: | Cesar Aybar <[email protected]> |
| License: | MIT + file LICENSE |
| Version: | 2026.9.27 |
| Built: | 2026-10-31 07:27:20 UTC |
| Source: | https://github.com/asterisk-labs/cozip |
Wraps stage_metadata() + arrow::write_parquet() +
stage_create(). For GeoParquet or custom write options, use
the staging functions directly.
create(out_path, table, temp_dir = NULL) write(out_path, table, temp_dir = NULL)create(out_path, table, temp_dir = NULL) write(out_path, table, temp_dir = NULL)
out_path |
Destination archive path. |
table |
arrow::Table with |
temp_dir |
Directory for the temp metadata parquet
(defaults to |
Absolute path of the created archive.
Reads the byte-0 index natively, fetches only the __metadata__
Parquet manifest, and decodes it with DuckDB, which loads no
extension. name, offset, size (and cozip:location when
location = TRUE) are always included; use columns to bring
extras, NULL brings all.
Archives using any profile other than Flat (profile = 1), including TACO
(profile = 2), are rejected.
read(source, columns = NULL, location = TRUE)read(source, columns = NULL, location = TRUE)
source |
Local path or http(s)/s3/gcs/azure/hf URL to the |
columns |
Character vector of extra columns. |
location |
Include the |
A tibble.
The metadata parquet is embedded verbatim, so any GeoParquet or schema metadata you wrote is preserved.
stage_create(out_path, paths, metadata_parquet, validate = TRUE)stage_create(out_path, paths, metadata_parquet, validate = TRUE)
out_path |
Destination archive path. |
paths |
data.frame with |
metadata_parquet |
Path to a parquet with |
validate |
If |
Absolute path of the created archive.
Computes offsets and sizes for the user entries and returns the
canonical metadata table plus a row-aligned paths data.frame.
The function checks each source and reads its file size; it does not read
payload contents.
stage_metadata(table)stage_metadata(table)
table |
arrow::Table with |
If you reorder metadata before writing it, reorder paths the
same way or stage_create() will reject the mismatch.
List with metadata (arrow::Table) and paths
(data.frame).