Indexing optimisation

[Julien B]

Hello

I have a table that contains thoses fields:

CREATE TABLE `test` (
`id_p` int(10) unsigned NOT NULL,
`id_s` int(10) unsigned default NULL,
`typ` enum('a','b') NOT NULL,
KEY `test_FKIndex1` (`id_s`),
KEY `test` (`id_p`),
KEY `index_test_1` (`typ`,`id_p`,`id_s`),
KEY `index_test_2` (`typ`,`id_p`),
CONSTRAINT `test_ibfk_1` FOREIGN KEY (`id_s`) REFERENCES `table_s` (`id_s`) ON DELETE NO ACTION ON UPDATE NO ACTION,
CONSTRAINT `test_ibfk_2` FOREIGN KEY (`id_p`) REFERENCES `table_p` (`id_p`) ON DELETE NO ACTION ON UPDATE NO ACTION
) ENGINE=InnoDB DEFAULT CHARSET=latin1

when i am doing this query:

select count(*) from test where id_p=<something> and typ=<something>

mysql does not look at index_test_2 index. (i saw it with the explain function).
If i force index_test_2 index with "use index(...)" syntax, it gives me the same row results in explain command.
It seems to be that mysql do not want to look at index_test_2 index.
Can someone explain me why ?

Is it beacuse i have 50% records with typ="a" and 50% with typ="b" ?

Thanks

Julien