• Declare a NOT NULL indexed column if possible. NULL requires additional server decisions and can be processed more simply and faster. • Avoid over indexing; don't index a column just because it can be indexed. If a column is never going to be referred to in comparisons, then there's no need to index it. • Every index created slows down table updates. If a row is inserted, an entry must be added to each of the table's indexes. • An index on a column that has very few distinct values is unlikely to do much good. • Choose unique and non-unique indexes appropriately. The choice may be influenced by the data type of a column and the level of distinctness. • Index a column prefix rather than the entire column. Shortening the length of the key values can improve performance by reducing the amount of disk I/O needed. • Avoid creating multiple indexes that overlap (have the same initial columns).
“The best way to be ready for the future is to invent it.”— John Sculley