If you run the pandas tutorial example, the dataframe you query has the same values as the dataframe written but with the index messed up (values around 01-01-1970 00:00) because the timestamps are badly computed.

This doesn't affect the version of influxdb-python you get by pip install, but the GitHub master branch (and my main pull request #388 at this time) are affected: the line protocol is broken.

There is no test to assert that the queried dataframe is equal to the written one.

I will try to make a specific PR for this issue when possible.