Earlier this year, I embarked on a bit of a project to tidy up the indexes in a sizeable database. This database has over 900 tables, and there are quite a few indexes which I have long suspected don’t need to exist for a variety of reasons. Some indexes are redundant, others don’t get used, still others could be combined into a single index. By reducing the number of indexes, we can improve write performance in the database, and the size of the database itself.