This question is a follow up to answered question (
http://forum.shephertz.com/?qa=4428/finddocumentbyid-vs-finddocumentbykeyvalue&show=4428#q4428).
Obviously, lookups by DocID will be significantly faster then by Key/Value pair, but can you go into more detail about how big of a performance difference there is? For instance, if I have a collection of one million json files, how long would it take for each query (by DocID and By Key/Value) to run? I am trying to identify the scope of the performance loss that could be incurred by using a Key/Value lookup.