（今天上网看到这篇博客，讲解索引的用处，比较基础，其实索引不止这三点）
I often see people confuse different ways MySQL can use indexing, getting wrong ideas on what query performance they should expect. There are 3 main ways how MySQL can use the indexes for query execution, which are not mutually exclusive, in fact some queries will use indexes for all 3 purposes listed here.
workarounds you can use though.
Using index to read data Some storage engines (MyISAM and Innodb included) can also use index to read the data, hence avoiding to read the row data itself. This is not simply savings of having 2 reads per index entry instead of one but it can save IO orders of magnitude in some cases – Indexes are sorted (at least on the page boundary) so doing index range scan you typically get many index entries from the same page but the rows itself can be scattered across many pages requiring potentially a lot of IOs. On top of that if you just need access to couple of columns
index can be simply much smaller than the data which is one of the reason covering indexes help to speed up queries even if data is in memory. If MySQL is only reading index and not accessing rows you will see “using index” in EXPLAIN output.
These are the main “core” use for indexes. You can also see others like using index for group by but I think they can be pretty much looked as one of these 3 ways described.
其实，在高并发的情况下，索引是一个非常好的解决并发问题的工具，虽然在mysql下体现得不是很明显，比如没有覆盖索引。
另外mysql也跟mssql一样支持index merge,不错
只是最近发觉mysql不支持并行查询，很是问题，如果在处理超大表的时候不并行，会有些问题。