Question

We have following scenario:

We have a chain of map-reduce processes implemented in java.Currently we are reading data from a mysql table and saving output to another mysql table.Now we may need data from another table as input to map/reduce process.

Possible Solutions:

a) Either we can have a join query for input to map process or

b) we can read needed data by making simple jdbc connection and requesting data again and again(although, i don’t prefer it).

Questions:

What are the best practices in such scenario?
We may move to mongoDB in future.What will be best practice in that scenario?

Answer 1

I think this is not possible for now.