list_layers
Parameters
Union[str, os.PathLike]
required
Path to one GeoPackage (
.gpkg) file, including Hadoop-supported
paths such as s3a://bucket/data.gpkg. Bytes and file-like objects
are not supported. A glob must resolve to exactly one file.Returns
pandas.DataFrame
Columns
name and geometry_type, sorted by name. Registered
nonspatial tables, including legacy aspatial tables, have None
as their geometry type. Raster tile tables are excluded.Unlike GeoPandas, this currently supports GeoPackage only. It uses
Sedona’s Spark reader, not Pyogrio. Only layer metadata is collected to
the driver; no feature geometries are read. The reader copies remote
GeoPackages in full to executor-local temporary storage, so remote
I/O and temporary disk space scale with the file size.Types describe declared metadata, including empty layers. Concrete
types permitting Z use a
Z suffix; M is not represented, matching
GeoPandas with Pyogrio.
Generic geometry layers report Unknown. Only core GeoPackage
geometry types are supported.Examples
points_from_xy
geopandas.points_from_xy, which returns an in-memory
GeometryArray, this scalable counterpart returns a distributed
GeoSeries so that its index and CRS metadata can be retained.
Parameters
iterable or pandas-on-Spark Series
required
Coordinate values. Distributed Series must come from the same frame;
numeric scalar coordinates are broadcast over that frame.
iterable or pandas-on-Spark Series
required
Coordinate values. Distributed Series must come from the same frame;
numeric scalar coordinates are broadcast over that frame.
iterable or pandas-on-Spark Series
default:"None"
Coordinate values. Distributed Series must come from the same frame;
numeric scalar coordinates are broadcast over that frame.
value
default:"None"
Coordinate Reference System accepted by
pyproj.CRS.from_user_input.Returns
GeoSeries
Distributed Point geometries. Shared distributed indexes, including
MultiIndexes and duplicate labels, are preserved.
Construction uses native
ST_Point or ST_PointZ expressions. It
does not collect pandas-on-Spark inputs or execute a Python row UDF.Local coordinates use NumPy broadcasting and are paired positionally, as
in GeoPandas. Local iterables, NumPy arrays, and pandas objects are
materialized on the driver and are intended for bounded inputs. Use
pandas-on-Spark Series for large coordinate columns.Distributed Series are paired within their shared Spark plan and retain
that plan’s index. Series from unrelated plans are rejected because they
have no safe distributed positional relationship.Examples
read_file
GeoDataFrame from a file.
Parameters
str
required
File path or file handle to read from. If the path is a directory,
Sedona will read all files in the directory into a dataframe.
Union[str, None]
default:"None"
The format of the file to read. If None, Sedona will infer the format
from the file extension. Note, inferring the format from the file extension
is not supported for directories.
Options:
- “shapefile”
- “geojson”
- “geopackage”
- “geoparquet”
GeoDataFrame.to_file: write GeoDataFrame to file
read_parquet
- if no geometry columns are read, this will raise a
ValueError- you should use the pandasread_parquetmethod instead.
Parameters
str, path object
required
list-like of strings
default:"None"
Not currently supported in Sedona
dict
default:"None"
Not currently supported in Sedona
tuple
default:"None"
Not currently supported in Sedona
dict
default:"None"
Not currently supported in Sedona
Returns
GeoDataFrame