We use a lot of SQL at my office but we want to move to HBase eventually since its the new wave of technology. But SQL has things like grouping and aggregation functions and equations that HBase doesn't, and we're looking into how we can put that functionality into HBase. I've looked around a lot but haven't found anything too useful to start my implementation. Are there any suggestions on how we could port SQL functions into HBase? Is it even possible?
1
-
3Moving to HBase just because "it's the new wave" is absolutely the wrong motivation. You should move because your current technology doesn't do what you want and HBase can deliver features you absolutely need.user330315– user3303152012-06-27 13:28:58 +00:00Commented Jun 27, 2012 at 13:28
Add a comment
|
5 Answers
Take a look to the open source project Phoenix.
Phoenix is a SQL skin over HBase, delivered as a client-embedded JDBC driver, powering the HBase use cases at Salesforce.com.
Comments
I'm not sure you are aware since you put the "vertica" tag in your post, but you can use both at the same time. Specifically, Vertica has a lot of Hadoop integration.
You should use the right tool or combination of tools for the job and not worry about using something because it is new.
Comments
Have a look at Apache Hive..It provides us SQL like interface on top of our Hadoop cluster. Along with that it also acts us a datawarehouse over the Hadoop cluster.
Comments
Take a look on this discussion : Group by In HBase It explains what are analytical capabilities of the HBase.
Comments
Use Phoenix SQL,It has fast performance due to use of server side co-processor for data aggregation, query parallelization for storing and accessing HBase database.
1 Comment
Matthew Haugen
In the future, please do not format things that are not code as code.