as per my understanding Shards are assigned to an Index. One index can have multiple Types holding number of Documents inside. Every Shard have Segments which actually store the Documents inside its data structures.
Now, my Question is that does a Shard stores documents of all the Types of the index in a mixed state? So, if I have 6 Types defined with each having 1 Million documents and I want to perform a query that is only related to 1 of the type documents, will the search cover all 6 Million documents to return the results?
If yes then how can I avoid this waste of resources? Should I build more individual indices instead of putting many types inside an index?