How to Use Apache Drill with Cassandra

Question

I am trying to query Cassandra using Apache Drill. The only connector I could find is here: http://www.confusedcoders.com/bigdata/apache-drill/sql-on-cassandra-querying-cassandra-via-apache-drill

However this does not build. It comes up with an artifact not found error. I also had another developer who is more versed in these tools take a stab at it, but he also had no luck.

I tried contacting the developer of the plugin I referenced, but the blog does not work and won't let me post comments. Has anyone got this plugin to work (if so how?) or is there another plugin or method I can use to connect apache drill to Cassandra? If anyone could show me how to connect an execute a simple SQL query that would be much appreciated.

Todhunter

You can give a change to prestoDB (from Facebook), to play SQL against Cassandra => prestodb.io. – Faviolafavonian

Answer 1

I looked at the latest Cassandra storage plugin patch and the latest apache drill source. The drill code has changed and the patch can no longer be applied. I then manually took the patch apart (it id mostly diff output). Most of the patch was new classes which I could easily add to the latest drill source tree. Most of the other updates were easy to insert into the current source. There were two specific classes that required some minor code modifications/extensions. I rebuilt the distribution from the modified source and installed the drill servers it on a 3 node cluster. The Cassandra schema failed to initialize properly throwing a null pointer exception one of the new classes. This leads me to believe that the (latest) modifed storage plugin is incompatible with the latest version of Cassandra. Since the author of the original storage plugin is unreachable and no one else is stepping up to support the code, this is a dead horse. Beat it if you must.

Toastmaster

Answer 2

I was the author of the patch written a year back. Could not get it merged into Drill then, and later got occupied with other stuffs :(

With so many changes to Drill internals, I am not sure what amount of welding would be needed at this point to get it working. Please use the code just as a reference for writing a Drill storage plugin.

Have added this banner on top of the blog post to save fellow developer's hours.

Costard

Answer 3

I don't know if anyone is still interested in this topic but I've been experimenting with this plugin and got it to work with Drill 1.18-SNAPSHOT. Here is a link to my branch with this code: 1. My plan is to submit this as a PR for Drill, but it still needs some work. This code will successfully query Cassandra 3.11.5 (latest stable version).

Outstand

Spark SQL works well with Cassandra, but would like to see Drill query Cassandra too. It would be simpler to use, battle tested JDBC/ODBC channels, and reusability of cache in case of multi-user environment. – Sulfide

Here is a link to the Drill pull request: github.com/apache/drill/pull/1960. This is not done yet, but if anyone wants to try it out, feedback is welcome. – Outstand