DataSourceError when writing GeoDataFrames to existing GeoPackage using Python

Question

How do I write many vector datasets (layers or GeoDataFrames) to a GeoPackage?

I have 12 GeoDataFrames that I want to store in one GeoPackage as separate layers. I'm writing these layers to the same file from multiple processes at the same time.

Below is the code I use, and it produces an error:

pyogrio.errors.DataSourceError: A file system object called '...
lines_1900.gpkg' already exists.

# write to GeoPackage
gpkg_name = 'lines_' + year + '.gpkg'
gpkg_file = os.path.join(out_folder, gpkg_name)
trip_lines.to_file(gpkg_file, layer=month, driver="GPKG")

Or is this a completely wrong approach?

Answer 1

I can run this code without errors, so I suggest your problem lies in a different part of your script.

This works with my QGIS-shipped python installation:

import geopandas as gpd
from shapely.geometry import Point, Polygon

# First GeoDataFrame: Points
gdf_points = gpd.GeoDataFrame({
    'name': ['A', 'B'],
    'geometry': [Point(0, 0), Point(1, 1)]
}, crs="EPSG:4326")

# Second GeoDataFrame: Polygons
gdf_polygons = gpd.GeoDataFrame({
    'id': [1],
    'geometry': [Polygon([(0, 0), (1, 0), (1, 1), (0, 1), (0, 0)])]
}, crs="EPSG:4326")

# Write both to the same GeoPackage file, each to a different layer
gdf_points.to_file("d:/tmp/example.gpkg", layer='points', driver="GPKG")
gdf_polygons.to_file("d:/tmp/example.gpkg", layer='polygons', driver="GPKG")

Answer 2

I cant reproduce your error.

Can you try this?

import geopandas as gpd
    
out_file = r"C:\Users\bera\Desktop\gistest\my_data.gpkg"
df = gpd.GeoDataFrame(data={"col1":[1]}, geometry=gpd.points_from_xy(x=[1], y=[1]), crs=3006)
    
#Write the dataframe as two different layers into the same geopackage
df.to_file(filename=out_file, layer="layer_1", driver="GPKG")
df.to_file(filename=out_file, layer="layer_2", driver="GPKG")

print(gpd.list_layers(out_file))
#       name geometry_type
# 0  layer_1         Point
# 1  layer_2         Point