Reconstruct. This should only be employed When you have a single dtype for all columns. Or else, you can end up with dtype item for all columns and converting them back again requires much more dictionary function. df.columns = new in which new may be the listing of new columns https://stack-ball-free-online-ga88642.frewwebs.com/37016485/5-easy-facts-about-stack-ball-game-for-kids-described