How do column oriented databases optimize the “select *” operation?

[presentationmall]

Clearly, a column oriented database is good for streaming through a single attribute over millions of records.But When we are filtering records based on the value of that attribute -- how do column oriented databases optimize the collection of attributes in the final output (i.e. "select 'name' from 'employee'" should be fast, but how does "select * from employee where name='jim'" get optimized under the hood?)