You could try reading it directly in Spark via JDBC, but for this size data you may want to do a higher throughput transfer into HDFS first and then load it from Spark.
On Wed, Oct 23, 2013 at 10:36 PM, Himanshu Bafna (HB) <[email protected]>wrote: > Hi > > I have use case where in our data is stored in Leveldb with data size > close to 100TB for 20 days. > Is it possible to integrate or load leveldb data into spark. > > OR should I first load the data from leveldb into HDFS for spark > consumption. > > Regards > Himanshu > >
