I'm performing some MySQL table maintenance that will mean removing some redundant columns and adding some new ones.
Some of the columns to drop are of the same type as ones to add. Would the procedure be faster if I took advantage of this and reused some of the existing columns?
My rationale is that changing column names should be a simple table metadata change, whereas removing and adding columns means either finding room at the end of the file (fragmenting data) or rebuilding every row with the correct columns so that they're at the same place on the disk.
The engine in question is MyISAM and I'm not up to scratch on how exactly it'll treat this so I'd like to hear from anyone who has been in the same situation before!