GeoDataFrames

Simple geographical vector interaction built on top of ArchGDAL. Inspiration from geopandas.

Some basic examples without explanation follow here, for a complete overview, please check the documentation.

Installation

]add GeoDataFrames

Usage

There's no special type here. You just use normal DataFrames with a Vector of ArchGDAL geometries as a column.

Reading

import GeoDataFrames as GDF
df = GDF.read("test_points.shp")
10×2 DataFrame
 Row │ geometry            name
     │ IGeometr…           String
─────┼────────────────────────────
   1 │ Geometry: wkbPoint  test
   2 │ Geometry: wkbPoint  test
   3 │ Geometry: wkbPoint  test
   4 │ Geometry: wkbPoint  test
   5 │ Geometry: wkbPoint  test
   6 │ Geometry: wkbPoint  test
   7 │ Geometry: wkbPoint  test
   8 │ Geometry: wkbPoint  test
   9 │ Geometry: wkbPoint  test
  10 │ Geometry: wkbPoint  test

You can also specify the layer index or layer name in opening, useful if there are multiple layers:

GDF.read("test_points.shp", 0)
GDF.read("test_points.shp", "test_points")

Any keywords arguments are passed on to the underlying ArchGDAL read function:

GDF.read("test.csv", options=["GEOM_POSSIBLE_NAMES=point,linestring", "KEEP_GEOM_COLUMNS=NO"])

Writing

using DataFrames

coords = zip(rand(10), rand(10))
df = DataFrame(geometry=createpoint.(coords), name="test");
GDF.write("test_points.shp", df)

You can also set options such as the layer_name, coordinate reference system, the driver and its options:

GDF.write("test_points.shp", df; layer_name="data", crs=EPSG(4326), driver="FlatGeoBuf", options=Dict("SPATIAL_INDEX"=>"YES"))

Note that any Tables.jl compatible table with GeoInterface.jl compatible geometries can be written by GeoDataFrames. You might want to pass which column(s) contain geometries, or by defining GeoInterface.geometrycolumns on your table. Multiple geometry columns, when enabled by the driver, can be provided in this way.

table = [(; geom=AG.createpoint(1.0, 2.0), name="test")]
GDF.write(tfn, table; geom_columns=(:geom,),)