Issue performing geometry operation in GeoJSON file "Error: Referenced column "geometry" not found in FROM clause"

[monasound]

Hi,

I am exploring Duckdb to work with spatial files.

I am trying to import GeoJSON files and perform ST_UNION. I can read the files without any issue, but when performing ST_UNION operation, it result in the following error:

An error occurred while performing ST_UNION: Binder Error: Referenced column "geometry" not found in FROM clause!
Candidate bindings: "geo_data_0.geom", "geo_data_0.Reference_Geography", "geo_data_0.ObjectId"
LINE 5: SELECT geometry FROM geo_data_0

1. Can someone help me explain the issue?
2. Looks like there is no compatibility in reading feature class of ESRI Geodatabase. Please confirm.

FYI: Somebody raised similar error a year ago. However, I have no idea whether it is updated or not. Please kindly share your comments

[Maxxen]

Hello!

1. What is the actual SQL query you're trying to run?
   As the error message explains, you're trying to pass a column named "geometry" to the ST_Union function, but there is no column with that name. Supposedly you got a table with a column named "geom" when you imported the geojson file.

2. Im not too familiar with ESRI Geodatabases, but what have you tried? I assume you've tried to pass a gdb file to st_read? Can you share the GDB file?

3. If there are other related issues it's generally a good idea to link them so that we know what you're talking about.

[monasound]

Below is the code I used.

1. Yes, there is a geometry column which I can see.
2. It is one of the geodatabase that I have downloaded and it opens correctly. What I encountered was error with respect to path. I am trying to fix this. If I have issues, I will share my code.

Thank you for your kind time.

`
import duckdb
import requests

geojson_urls = [
'https://services-eu1.arcgis.com/BuS9rtTsYEV5C0xh/arcgis/rest/services/CensusHub2022_T1_2_LEA/FeatureServer/0/query?outFields=&where=1%3D1&f=geojson',
'https://services-eu1.arcgis.com/BuS9rtTsYEV5C0xh/ArcGIS/rest/services/grid_population_2022_view/FeatureServer/0/query?outFields=&where=1%3D1&f=geojson' # Replace with your second URL
]

con = duckdb.connect()

con.execute("INSTALL spatial;")
con.execute("LOAD spatial;")

for index, geojson_url in enumerate(geojson_urls):

response = requests.get(geojson_url)
response.raise_for_status()  # Raise an error for bad responses


if 'application/json' not in response.headers.get('Content-Type', ''):
    raise ValueError(f"The URL {geojson_url} did not return a valid GeoJSON file.")


local_geojson_path = f'data_{index}.geojson'
with open(local_geojson_path, 'w', encoding='utf-8') as f:
    f.write(response.text)


try:
    query = f'''
    CREATE TABLE geo_data_{index} AS SELECT * FROM ST_Read('{local_geojson_path}');
    '''
    
    con.execute(query)

except Exception as e:
    print(f"An error occurred while reading the GeoJSON file from {local_geojson_path}: {e}")

try:
union_query = '''
CREATE TABLE geo_data_union AS
SELECT ST_Union(geometry) AS geometry
FROM (
SELECT geometry FROM geo_data_0
UNION ALL
SELECT geometry FROM geo_data_1
);
'''

con.execute(union_query)


union_result = con.execute("SELECT * FROM geo_data_union").fetchdf()


print("Union of data_0 and data_1:")
print(union_result)

except Exception as e:
print(f"An error occurred while performing ST_UNION: {e}")

con.close()`

[Maxxen]

Geometry is the name of the type not the column. The geometry column you get from ST_Read is called "geom", not "geometry"