I have a DataFrame where the geometry data was imported from an Oracle database. It looks like this: df_wkt.head() The output is: FID WKT_GEOM 0 214213 CURVEPOLYGON (COMPOUNDCURVE ((2694062.785 1263... 1 214214 CURVEPOLYGON (COMPOUNDCURVE (CIRCULARSTRING (2... 2 214209 CURVEPOLYGON (COMPOUNDCURVE ((2700414.355 1262... 3 214210 CURVEPOLYGON (COMPOUNDCURVE (CIRCULARSTRING (2... 4 214211 CURVEPOLYGON (COMPOUNDCURVE ((2696513.166 1263... A single object looks like this: df_wkt["WKT_GEOM"].iloc[1]: <oracledb.LOB at 0x7244f423ca70> I want to convert these geometries (currently of type <oracledb.LOB>) into a format that is compatible with a PostGIS database. There are two main challenges: Converting the Oracle LOB objects into usable geometries. Converting the CURVEPOLYGON geometries into linear polygons, as I need the final output to contain only linear (non-curved) geometries. Continue reading...