write_catalog

Catalog.write_catalog(base_catalog_path: str | Path | UPath, *, catalog_name: str | None = None, default_columns: list[str] | None = None, inherit_provenance: bool = False, addl_hats_properties: dict | None = None, as_collection: bool = True, overwrite: bool = False, resume: bool = False, progress_bar: bool = True, tqdm_kwargs: dict | None = None, create_parquet_metadata: bool = True, create_thumbnail: bool = False, create_per_partition_statistics: bool = True, error_if_empty: bool = True, create_summary: bool = False, should_write_skymap: bool = True, skymap_alt_orders: list[int] | None = None, npix_suffix: str = '.parquet', npix_parquet_name: str | None = None, write_table_kwargs: dict | None = None)

Save the catalog to disk in HATS format.

Parameters:

base_catalog_pathstr | Path | UPath,

Location where catalog is saved to

catalog_namestr

The name of the catalog to be saved

default_columnslist[str]

A metadata property with the list of the columns in the catalog to be loaded by default. By default, uses the default columns from the original hats catalog if they exist.

inherit_provenancebool, default False

If the original catalog had some provenance info (like creator or bib references), should this new catalog retain all of it?

addl_hats_propertiesdict or None, default None

key-value pairs of additional properties to write in the hats.properties file.

as_collectionbool, default True

If True, saves the catalog and its margin as a collection

overwritebool, default False

If True existing catalog is overwritten

resumebool, default False

If True, resumes a previous write operation, skipping partitions that were already written to disk.

progress_barbool, default True

If True, shows a progress bar for the export process. Defaults to True.

tqdm_kwargsdict, default None

Additional kwargs to pass to the tqdm progress bar.

create_parquet_metadatabool, default True

Should we create /dataset/_metadata parquet from all data partitions.

create_thumbnailbool, default False

If True, create a data thumbnail of the catalog for previewing purposes. Defaults to False.

create_per_partition_statisticsbool, default True

Should we create per_partition_statistics.parquet, based on footers from all data partitions

error_if_emptybool, default True

If True, raises an error if the output catalog is empty

create_summarybool, default False

If True, writes README.md summary file(s) describing the catalog. When as_collection is True, this generates a summary for the collection, the main catalog, and the margin (if it exists).

should_write_skymap: bool, default True

Should we write a skymap.fits file, with the point distribution of the catalog? Main catalogs should contain skymap fits files, generally.

skymap_alt_orderslist[int] or None, default None

We will write a skymap file at the histogram_order, but can also write down-sampled skymaps, for easier previewing of the data.

npix_suffixstr, default ‘.parquet’

Extension for the parquet file (or / if a directory)

npix_parquet_namestr | None, default None

Name of the pixel parquet file to be used when npix_suffix=/. By default, it will be named after the pixel with a .parquet extension (e.g. ‘Npix=10.parquet’).

write_table_kwargs: dict or None, default None

Arguments to pass to the parquet write operations

Examples

Write a small synthetic catalog to disk:

>>> import lsdb
>>> from lsdb import generate_data
>>> nf = generate_data(1000, 5, seed=0, ra_range=(0.0, 300.0), dec_range=(-50.0, 50.0))
>>> catalog = lsdb.from_dataframe(nf.compute()[["ra", "dec", "id"]], catalog_name="demo")
>>> catalog.write_catalog(<your path here> / "demo_catalog", overwrite=True)