Migration from old tables broken if they have more than one row

[nhoening]

Saw this on the staging server which tried to upgrade from an older state:

remote:   File "/home/seita/flexmeasures-staging/flexmeasures-work/flexmeasures/data/migrations/versions/ad98460751d9_remove_obsolete_tables.py", line 42, in upgrade        
remote:     result = db.session.execute(        
remote:   File "/home/seita/.virtualenvs/bvp-staging-venv/lib/python3.8/site-packages/sqlalchemy/engine/result.py", line 1487, in scalar_one_or_none        
remote:     return self._only_one_row(        
remote:   File "/home/seita/.virtualenvs/bvp-staging-venv/lib/python3.8/site-packages/sqlalchemy/engine/result.py", line 805, in _only_one_row        
remote:     raise exc.MultipleResultsFound(        
remote: sqlalchemy.exc.MultipleResultsFound: Multiple rows were found when one or none was required    

This code doesn't seem to be able to deal with multiple rows, and was not tested well:

           if inspect.has_table(table):
                result = db.session.execute(
                    sa.text(f"SELECT 1 FROM {table};")
                ).scalar_one_or_none()
                if result:
                    tables_with_data.append(table)
            else:
                print(f"Table {table} not found, continuing...")

...

    if tables_with_data:
        click.confirm(
            f"The following tables still have data and will be dropped by this upgrade: {tables_with_data}. Use `flexmeasures db-ops dump` to create a backup. Are you sure you want to upgrade the database?: ",
            abort=True,
        )

This migration is over a year old, so the urgency is not super-high, as there are probably no other old databases around at this time.