I have a big MySQL table (around 50 GB, tens of millions of records), with one index key (a foreign key) that has around 2000 possible d开发者_开发百科ifferent values, and another index key (a VARCHAR(2)) that has 3 possible values.
Can I expect a significant performance increase if I partition the table based on either (or both) indices?
You can't partition on a column that is not an index (or part of it), so yeah... it is worth it.
Have you thought of de-normalising the table when there is three possible values if you have performance problems?
精彩评论