In cassandra, it\'s well known that secondary indexes should be used very sparingly.
If I have a table for example:
User(username, usertype, email, etc..
The good option here would be to create a composite primary key consisting of username and usertype with username being partition key and usertype a cluster key. You will not even need an index and the query will work.
CREATE TABLE users (
username text,
usertype text,
....
PRIMARY KEY ((username), usertype)
)