At the language level, yes. At the on-computers-all-over-the-world-on-petabyte-datasets level? No. That was Google.

That's like those 'on the internet' patents then. Only this time it is the 'on a cluster' variation. Take an existing scheme, implement it on a cluster, instant market lock.

Yeah, it's not that simple. You're talking about one of the biggest scaling problems ever addressed. You have to figure out how to shard the data, how to deal with failures and inconsistencies, how to return results from terabytes of data, and how to make it all fast enough that your search results come back in less than a second.

Map-reduce isn't used in search queries only for indexing.