How to map tables from rdbms and hbase

2
votes

In my current project, I have been given task to design HBase tables from a set of RDBMS tables with E-R diagrams. 1) What can be the approach to design HBase tables by looking into the RDBMS tables? I am trying to find out some guidelines for this.

2) And also, while going through HBase material, I have come across row-oriented and column-oriented concepts. Which one is good?

3) Is Column Family a must for HBase? If yes, then how to decide on a column family, after looking at a rdbms table? I assume, that, columns which are changing constantly, should be "values" and can be part of column families, correct?

Regards, Shouvanik Big Data Developer

1
hbasemultiple-columns
see stackoverflow.com/questions/15218672/… – Arnon Rotem-Gal-Oz

1 Answers

1
votes

Well I've been through the same process the last few weeks and from all the resources I found these were the most interesting. I feel like they can definitely help you.

A very clarifying article about HBase and it's basic concepts:

http://jimbojw.com/wiki/index.php?title=Understanding_Hbase_and_BigTable

A video from last year's HBaseCon about how to design a schema in HBase opposed to traditional databases:

http://www.cloudera.com/content/cloudera/en/resources/library/hbasecon/video-hbasecon-2012-hbasecon-2012.html