NoSQL Database and Reporting

Question

I am at the stage of the architecture of an academic project involving billions of entries. The project should be very light in terms of computing power and high scalability. The structure of the information is very simple: I need to keep a list of elements, each of which has different functions. Feature are integers, decimals, dates, strings, etc. When data is imported, function types are known. In addition, functions can be used to refer to other elements.
I need to get and sort a list of elements by its functions (more than one) - possibly using queries such as>, <, = and regular expressions, length, left, right, middle for strings between function values ​​and against arbitrary user input.

Reporting in the sense of sums, averages, groupings is also necessary because of requirements that are more relaxed - there is no need for the full capabilities of the cube, but better.

I am very new to the world of NoSQL. What would you suggest?.

Answer 1

If you study the tutorials for MongoDB, they have, in my opinion, the best introduction to the Map / Reduce system, which is used for querying and aggregating.

I really wonder why you agreed in advance that NoSQL is a route. Although different elements may have different schemes, is there a fixed number of entities and attributes, and why have you (if any) excluded SQL, which, after all, has decades of accumulated functions for storing and querying data.

Answer 2

, , , .

, Apache Pig and Hive. adhoc-, .