How to organize a quick search through 3 billion documents?

Question

MongoDB

Hello
Life pushed me with the need for a quick search among 3 billion documents. Roughly speaking, this is a list of key queries from PS users, from which you need to select those that include, for example, the word "download".
I uploaded a demo billion into MongoDb (by the way, I received 239 GB of data in 139 files) and tried to make a selection through Matches (I work with Mongo through the c # driver).
Even the simple Count of this query took so long that I didn't have the patience. At the same time, 10% of the processor and a bit of memory were occupied by the mongo.
And I would like all this to be done in seconds, not minutes or hours.
Tell me, in which direction to dig?
Thank you.

Answer 1

elasticsearch

Answer 2

Towards mongo index or text search engines - lucene, sphinx..

Answer 3

Roughly speaking, this is a list of key queries from PS users, from which you need to select those that include, for example, the word "download".

If the list of keywords is not large, then you can make an indexed array in the documents and put these keywords (or their identifiers) there.
Otherwise use sphinx/elasticsearch. You can also use text index from mongoDB, but it seemed to me too gluttonous.

I tried to make a selection through Matches

At the same time, it enumerates and checks all documents, which is why it takes so long.