Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
Rework hdf output #11
orca.run( ['my_step'], range(2010, 2030), data_out=out_file, out_interval=5, out_base_tables=orca.list_tables(), out_run_tables=['a_table'] )
# just output a couple tables orca.write_tables(out_h5, ['buildings', 'persons']) # write out everything that's registered orca.write_tables(out_h5) # still allow for step inference step_tables = orca.get_step_table_names(['my_step']) orca.write_tables(out_h5, step_tables)
The docstrings and tests have been updated to reflect these changes.
I looked through the repo and the only method calling write_tables is the run method. So all tests pass fine. However, if there are calls to orca.write_tables coming from outside orca, then these will now fail. I don't see why this previously would be called outside the context of a run, but if that is the case then I can revert the signature back to the previous version.
This passes for python 3.3 but fails for 2.7 and 3.4 on an assert column series equal.
It appears that setting values via .loc is changing the data type from int64 to float. Should the update_col_from_series method be updated so that the data type of the incoming series is cast to match the local column?
Or should update_col_from_series check the data types of the series and the incoming values and raise an exception if they don't match?