How can hive run an mapreduce algorithm?

Question

Now I'm implementing apache hive, and I now that hive run an mapreduce algorithm too like the usual hadoop mapreduce. So, I want to know, about an explanation about mapreduce, mapper and reducer from a hive query. How can that's query do an mapreduce?

Thanks all.....

Answer 1

Similar to what Oracle or MySQL does with SQL when you run an SQL query in those DB's. There is a parser which parses the query and then builds out Oracle/MySQL specific directives which the DB engine understands. Similarly Hive has a parser & optimizer which converts SQL query to MR.

For eg select count(*) from table_name; Would translate to Mapper (Read file and count) --> Reducer (sum all counts and return).