The simplest way is to just use thrift batch_mutate. If Cassandra CPU is your bottleneck then using the binary load method from StorageProxy can help (see contrib/bmt_example).
If Casssandra disk or network is your bottleneck then binary load won't really help. On Thu, May 6, 2010 at 7:51 AM, Dop Sun <su...@dopsun.com> wrote: > Hi, > > > > I just discovered that the json file exported by sstable2json contains more > than the data itself, like deletedAt values. > > > > I’m thinking whether there is a tool can import some initial data? > > > > When we are doing the typical RDBMS system, this is how we are doing: > > 1) Define the schema > > 2) Initialize the database with initial data using Insert statements > > 3) Running the application > > > > If following these steps in Cassandra: > > 1) Define the schema: this is done by define the KeySpace > > 2) Initialize the data with initial data: This is currently not there. > We can import data exported from existing installation, but not from nothing > > 3) Running the application > > > > Thanks, > > Regards, > > Dop -- Jonathan Ellis Project Chair, Apache Cassandra co-founder of Riptano, the source for professional Cassandra support http://riptano.com
