Inconsistent default geometry column name for postgis

[ondave]

It's not a major issue, but I am wondering why the default geometry column name is "geom" for read_postgis? To create a GeoDataFrame, a "geometry" column must be supplied. However the default column name for read_postgis is "geom". So combining a write and read with default arguments fails:

import geopandas
from shapely.geometry import Point
from sqlalchemy import create_engine

URI="postgresql:///user:password@localhost:5432/test"
con=create_engine(URI)

d = {'col1': ['name1', 'name2'], 'geometry': [Point(1, 2), Point(2, 1)]}
gdf = geopandas.GeoDataFrame(d, crs="EPSG:4326")

gdf.to_postgis("my_table",con)
gdf_read=geopandas.read_postgis("SELECT * FROM my_table",con)

Exception has occurred: ValueError
Query missing geometry column 'geom'

[martinfleis]

PostGIS default name for a geometry column is geom. Since geopandas uses geometry as a default, it causes this inconsistency.

If you read PostGIS table which is not created by geopandas, the default in most cases works as it is (geom), that is why it is there.

We may consider changing default in to_postgis which would rename geometry to geom but I am not entirely sure about that.