Index structure

[Shailendra mudgal]

Hi,

I have a table of two columns.
CREATE TABLE `tab1` (
`col1` int(11) NOT NULL,
`col2` varchar(20) NOT NULL DEFAULT '',
PRIMARY KEY (`col1`,`col2`),
KEY `col2_index` (`col2`),
KEY `col1_index` (`col1`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8;


This table may contain 20 Million rows.
The queries to this table are doing self join eg:
select distinct e1.col1 from tab1 e1, tab1 e2 where (e1.col2='<somevalue1>') and (e2.col2='<somevalue2>') and e1.col1=e2.col1;

These queries are taking a lot of time (around 40 sec).

Is there a way by which I can optimize the structure of table OR query so that it executes in minimum amount of time.

Thanks in advance...