(changing to user hbase distlist instead of general hadoop) Hive runs as MapReduce jobs. If you are looking for "quick" (i.e., non-MapReduce) access to tables, you need to use the HBase client (Get, Scan).
http://hbase.apache.org/book.html#datamodel -----Original Message----- From: Mag Gam [mailto:[email protected]] Sent: Sunday, April 10, 2011 11:37 AM To: [email protected] Cc: Ted Yu Subject: Re: hbase It seems with HIVE is a large latency. Is there anything else out there which has less latencies? On Sun, Apr 10, 2011 at 9:56 AM, Ted Yu <[email protected]> wrote: > Mapreduce isn't required. > For your query, Hive would be a better fit. > > On Sun, Apr 10, 2011 at 5:36 AM, Mag Gam <[email protected]> wrote: > >> Just curious, does hbase require mapreduce? Basically, I have several >> terabytes of data and I would like to query it similar to sql fashion. >> Was wondering if mapreduce was required. >> >
