The following code does not change the scores dataframe. scores[["Quiz1", "Quiz2"]] = scores.[["Quiz1", "Quiz2"]].bfill(axis=1) However, this statement does the change: scores.loc[:, ["Quiz1", "Quiz2"]] = scores.loc[:, ["Quiz1", "Quiz2"]].bfill(axis=1) And interestingly, the first approach does change the scores in the following statement: scores[['Quiz1', 'Quiz2', 'Quiz3']] = scores[['Quiz1', 'Quiz2', 'Quiz3']].astype('int64') I think this is about working on the copy or view of the dataframe. But, I feel like there is an inconsistency how pandas dataframe behave. Can someone explain what is happing here? Continue reading...
