Reconstruct. This could only be applied if you have only one dtype for all columns. Otherwise, you will end up with dtype item for all columns and converting them again involves extra dictionary perform. df.columns = new exactly where new may be the list of new columns names is so https://trumanl396hdx6.mysticwiki.com/user