Tom, Thanks, taking a look at actual examples would help us a lot.
One of the concerns we have Thrift is how the writes will get distributed across the Region Servers, this is still a new paradigm for us moving away from MySQL, so any insights on that would be great. thanks -srinivas On Fri, Apr 27, 2012 at 10:36 AM, Thomas Macke <[email protected]> wrote: > I have some simple C++ thrift codes that create & delete tables, show the > status of all tables in an hbase, a simple delim sep line inserter and > simple scanner. These are not production codes, but I have craeted > multi-million line tables with them. I can send them to you if you'd like > to see some real thrift examples. > > Tom > > On 4/26/12 7:47 PM, "Srinivas Krishnan" <[email protected]> wrote: > >>We are transitioning from a BerkeleyDB+MySQL backend to testing Hbase. >>One of the toughest challenges we are facing is porting the interface >>over. >> >>All our clients were using C/C++ APIs of Berkeley and MySQL to write >>data, is there a standard way to do that for HBase. The current method >>we are exploring is using JNI to interface with the Java connector, >>but its more trouble than it worth. >> >>Is there a C++ client library for HBase, or is the expected mechanism >>Thrift ? >> >>One point to mention we have used Avro in the past to serialized some >>data, would that be preferred over Thrift ? >> >>-sri >
