search

GregorCH / ipet

Interactive Performance Evaluation Tools for Optimization Software
MIT License
26 stars 6 forks source link

fix bug when comparing nonexistent columns #62

Closed fschloesser closed 6 years ago

fschloesser commented 6 years ago

We talked about this. It seems pandas can cope better with nonexisting data.

GregorCH commented 6 years ago

Hm, I am not convinced. These concat-operations look unnecessarily expensive (they create a new data frame object that gets directly reduced afterwards) to me. Usually, I would say, what does it matter, but currently, we try to make it faster.

The fastest thing to do would be to say: "I do not know what this minimum/maximum value is supposed to be, because it is not a column, and not an integer or float, either."

I would vote for printing a warning in this case and continue, because errors because some data is missing are hard to fix from a user's point of view.

fschloesser commented 6 years ago

Okay, so what would you do, let ipet crash with that warning/error or somehow skip the column in question?