Python tools for geographic data
Small bug-fix release with several fixes for compatibility with the upcoming pandas 3.0, numpy 2.0 and fiona 1.10 releases.
Notes on dependencies:
API changes:
unary_union is now deprecated and replaced by the union_all method (#3007).align keyword in binary methods now defaults to None, treated as True. Explicit True
will silence the warning about mismachted indices. (#3212)sjoin method will now preserve the name of the index of the right
GeoDataFrame, if it has one, instead of always using "index_right" as the
name for the resulting column in the return value (#846, #2144).New methods:
line_merge method from shapely to GeoSeries/GeoDataframe (#3214).set_precision and get_precision methods from shapely to GeoSeries/GeoDataframe (#3175).count_coordinates method from shapely to GeoSeries/GeoDataframe (#3026).minimum_clearance method from shapely to GeoSeries/GeoDataframe (#2989).is_ccw method from shapely to GeoSeries/GeoDataframe (#3027).is_closed attribute from shapely to GeoSeries/GeoDataframe (#3092).force_2d and force_3d methods from shapely to GeoSeries/GeoDataframe (#3090).contains_properly method from shapely to GeoSeries/GeoDataframe (#3105).snap method from shapely to GeoSeries/GeoDataframe (#3086).transform method from shapely to GeoSeries/GeoDataFrame (#3075).dwithin method to check for a "distance within" predicate on
GeoSeries/GeoDataFrame (#3153).to_geo_dict method to generate GeoJSON-like dictionary from a GeoDataFrame (#3132).New features and improvements:
predicate="dwithin" option and distance argument to the sindex.query() method
and sjoin (#2882).__repr__ now trims trailing zeros for a more readable
output (#3087).on_invalid parameter to from_wkt and from_wkb (#3110).make_valid option in overlay now uses the make_valid method instead of
buffer(0) (#3113)."geometry" as dtype to pd.read_csv will now return a GeoSeries for
the specified columns (#3101).read_file for the mask keyword for the pyogrio engine (#3062).read_file for the columns keyword for the fiona engine (#3133).show_bbox, drop_id and to_wgs84 arguments to allow further customization of
GeoSeries.to_json (#3226)Backwards incompatible API changes:
explode(.., index_parts=True) is now
set to false for consistency with pandas (#3174).set_geometry has been changed when passed a (Geo)Series ser with a name.
The new active geometry column name in this case will be ser.name, if not None, rather than
the previous active geometry column name. This means that if the new and old names are
different, then both columns will be preserved in the GeoDataFrame. To replicate the previous
behaviour, you can instead call gdf.set_geometry(ser.rename(gdf.active_geometry_name)) (#3237).Potentially breaking changes:
read_file
now returns a Pandas DataFrame instead of a GeoDataFrame with an empty
geometry column.Bug fixes:
GeoDataFrame.merge() incorrectly returning a DataFrame instead of a
GeoDataFrame when the suffixes argument is applied to the active
geometry column (#2933).GeoDataFrame constructor where if geometry is given a named
GeoSeries the name was not used as the active geometry column name (#3237).Deprecations and compatibility notes:
The deprecation of geopandas.datasets has been enforced and the module has been
removed. New sample datasets are now available in the
geodatasets package (#3084).
Many longstanding deprecated functions, methods and properties have been removed (#3174), (#3189)
geopandas.io.read_file, geopandas.io.to_file and geopandas.io.sql.read_postgis.
geopandas.read_file, geopandas.read_postgis and the GeoDataFrame/GeoSeries to_file(..)
method should be used instead.GeometryArray.data property, np.asarray(..) or the to_numpy()
method should be used instead.sindex.query_bulk method, using sindex.query instead.sjoin parameter op, predicate should be supplied instead.__xor__, __or__, __and__ and
__sub__. Instead use methods symmetric_difference, union, intersection and
difference respectively.plot_polygon_collection,
plot_linestring_collection and plot_point_collection, use the GeoSeries/GeoDataFrame .plot
method directly instead..plot parameters axes and colormap, instead use
ax and cmap respectively.Fixes for compatibility with psycopg (#3167).
The include_fields and ignore_fields keywords in read_file() are deprecated
for the default pyogrio engine. Currently those are translated to the columns keyword
for backwards compatibility, but you should directly use the columns keyword instead
to select which columns to read (#3133).
The drop keyword in set_geometry has been deprecated, and in future the drop=True
behaviour will be removed (#3237). To prepare for this change, you should remove any explicit
drop=False calls in your code (the default behaviour already is the same as drop=False).
To replicate the previous drop=True behaviour you should replace
gdf.set_geometry(new_geo_col, drop=True) with
geo_col_name = gdf.active_geometry_name
gdf.set_geometry(new_geo_col).drop(columns=geo_col_name).rename_geometry(geo_col_name)
Small bug-fix release:
pandas.concat CRS consistency checking where CRS differing by WKT whitespace only were treated as incompatible (#3023).Small bug-fix release:
overlay where using buffer(0) instead of make_valid internally produced invalid results (#3074).explore() method when the active geometry contains missing and empty geometries (#3094).Small bug-fix release updating the GeoParquet spec version and with a critical security fix:
read_parquet and read_feather for CVE-2023-47248 (#3070).GeoPandas will use Shapely 2.0 by default instead of PyGEOS when both Shapely >= 2.0 and PyGEOS are installed. PyGEOS will continue to be used by default when PyGEOS is installed alongside Shapely < 2.0. Support for PyGEOS and Shapely < 2.0 will be removed in GeoPandas 1.0. (#2999)
API changes:
seed keyword in sample_points is deprecated. Use rng instead. (#2913).New methods:
concave_hull method from shapely to GeoSeries/GeoDataframe (#2903).delaunay_triangles method from shapely to GeoSeries/GeoDataframe (#2907).extract_unique_points method from shapely to GeoSeries/GeoDataframe (#2915).frechet_distance() method from shapely to GeoSeries/GeoDataframe (#2929).hausdorff_distance method from shapely to GeoSeries/GeoDataframe (#2909).minimum_rotated_rectangle method from shapely to GeoSeries/GeoDataframe (#2541).offset_curve method from shapely to GeoSeries/GeoDataframe (#2902).remove_repeated_points method from shapely to GeoSeries/GeoDataframe (#2940).reverse method from shapely to GeoSeries/GeoDataframe (#2988).segmentize method from shapely to GeoSeries/GeoDataFrame (#2910).shortest_line method from shapely to GeoSeries/GeoDataframe (#2960).New features and improvements:
exclusive parameter to sjoin_nearest method for Shapely >= 2.0 (#2877)to_file() method will now automatically detect the FlatGeoBuf driver
for files with the .fgb extension (#2958)Bug fixes:
"crs" (#2944)explore when using UserDefined bins (#2923)apply with axis=1 where the given user defined function returns nested
data in the geometry column (#2959)np.int32 and pd.Int32Dtype columns (#2950)assert_geodataframe_equal now handles GeoDataFrames with no active geometry (#2498)Notes on (optional) dependencies:
Deprecations and compatibility notes:
geom_almost_equals() methods have been deprecated and
geom_equals_exact() should be used instead (#2604).Acknowledgments
Thanks to everyone who contributed to this release! A total of 16 people contributed patches to this release. People with a "+" by their names contributed for the first time.
Full Changelog: https://github.com/geopandas/geopandas/compare/v0.13.2...v0.14.0
Small bug-fix release with this critical fix:
file://..) using geopandas.read_file (#2948).Small bug-fix release with this critical fix:
geopandas.read_file (#2908). This
restores the behaviour to download all data up-front before passing it to the
underlying engine (fiona or pyogrio), except if the server supports partial requests
(to support reading a subset of a large file).New methods:
sample_points method to sample random points from Polygon or LineString geometries (#2860).hilbert_distance() method that calculates the distance along a Hilbert curve for each geometry in a GeoSeries/GeoDataFrame (#2297).sort_values()) based on the distance along the Hilbert curve (#2070).get_coordinates() method from shapely to GeoSeries/GeoDataframe (#2624).minimum_bounding_circle() method from shapely to GeoSeries/GeoDataframe (#2621).minimum_bounding_radius() as GeoSeries method (#2827).Other new features and improvements:
GeoSeries.fillna via another GeoSeries (#2535).min_zoom and max_zoom inside the map_kwds argument for .explore() (#2599).mode="a" or append=True) in to_file() using engine="pyogrio" (#2788).to_wgs84 keyword to to_json allowing automatic re-projecting to follow the 2016 GeoJSON specification (#416).to_json output now includes a "crs" field if the CRS is not the default WGS84 (#1774).geometry attribute of GeoDataFrame without an active geometry column related to the default name "geometry" being provided in the constructor (#2577)Deprecations and compatibility notes:
unary_union will return 'GEOMETRYCOLLECTION EMPTY' instead of None for all-None GeoSeries. (#2618)query_bulk() method of the spatial index .sindex property is deprecated in favor of query() (#2823).Bug fixes:
plot() if an empty or missing geometry is present (#2224)explore() (#2657)to_parquet/to_feather to not write an invalid bbox (with NaNs) in the metadata in case of an empty GeoDataFrame (#2653)to_parquet/to_feather to use correct WKB flavor for 3D geometries (#2654)read_file to avoid reading all file bytes prior to calling Fiona or Pyogrio if provided a URL as input (#2796)copy() downcasting GeoDataFrames without an active geometry column to a DataFrame (#2775)iterfeatures() method of GeoDataFrame to correctly handle non-scalar values when na='drop' is specified (#2811)plot (#2886)Notes on (optional) dependencies:
Acknowledgments
Thanks to everyone who contributed to this release! A total of 32 people contributed patches to this release. People with a "+" by their names contributed a patch for the first time.
Bug fixes:
to_crs() when using PyGEOS or
Shapely >= 2.0 (previously the z coordinates were lost) (#1345).naturalearth_lowres built-in dataset (#2670)