0

Is there a column store similar to Vertica that is built on top of Hadoop.. I am not talking about HBase as it is sparse matrix store and can not get the level of compression that can be achieved by something like Vertica?

Thanks

StackUnderflow
  • 24,080
  • 14
  • 54
  • 77

4 Answers4

3

Are you looking for something like RCFile? It is a file type that uses a columnar store internally.

Steve Severance
  • 6,611
  • 1
  • 33
  • 44
1

RCFile is a good start. RCFile stores data in a PAX layout -- columnar within blocks that could be as large as HDFS's block size. There is a paper at VLDB 2011 describing another columnar storage format here and a blog post with a short comparison to RCFile here.

Sandeep
  • 11
  • 1
0

Look on Hadapt http://hadapt.com/

This is a commercial version of HadoopDB http://db.cs.yale.edu/hadoopdb/hadoopdb.html developed at Yale University. It can work with a column-oriented DB (which is installed on every node of a computer cluster) while leverages Hadoop for fault-tolerant execution.

Antonio
  • 756
  • 7
  • 26
0

I haven't worked with Hadoop, but I know Vertica has been trying to integrate with Hadoop.

http://www.vertica.com/the-analytics-platform/native-bi-etl-and-hadoop-mapreduce-integration/

geoffrobinson
  • 1,580
  • 3
  • 15
  • 23