Checkout differentiators between thrift n cql3. Storage engine is same but
they differs in metadata. Prior to 2.0 it was mix of both. I would still
suggest the same. Checkout earlier threads around map collection support,
bulk loading, dynamic column support comparison b/w thrift and cql3. Ease
of
Stick with cql3 going forward. Last i checked, there is no further dev on
thrift. I had worked with Thrift based c* api for 2 years in Python based
pycassa and experience was not very satisfactory. I had not done comparisions
between pycassa and cql so cant say. Cql isvery simple any way.—
Sent
Hi every one,
I'm newcomer to Cassandra, so I would like to know about performance
between Thrift (Hector) vs CQL3, specially about the speed (Thrift based
on RPC, CQL3 based on binary protocol).
Currently I'm using Cassandra 1.2 , which version CQL3 of
JavaDriver-Datastax is stable for it?