Comments on: WTF is a SuperColumn? An Intro to the Cassandra Data Model

By: Brian Mansell

Brian Mansell — Fri, 05 Mar 2010 03:21:02 +0000

First of all, I must say that this is an excellent article!

As this appears as one of the first Google results when searching for Cassandra, I’d recommend posting an update which reflects the storage configuration changes (e.g. type -> ColumnType) made in recent versions of Cassandra.

thanks for all your efforts with this tutorial!

By: Pablo Rosales

Pablo Rosales — Thu, 04 Mar 2010 19:40:19 +0000

Hey, great post, its very useful! At first Cassandra is confusing but so far I love it.

By: Marcin

Marcin — Tue, 02 Mar 2010 12:21:49 +0000

I’m just trying to get my head around Cassandra, and I _think_ I understand how the data model works. There is however one simple use case I can’t work out how you’d represnent: for the AddressBook example, say you had two friends with the same real full name – how would you key the super-columns then? Obviously you’d need a unique identifier (say a UUID), but what if you wanted to then get back a list of your friends in alphabetical order of real name?

By: Yonatan

Yonatan — Thu, 25 Feb 2010 15:08:59 +0000

and another one – I understand that I need to work hard cause the data is not normalized – so when data is changed I need manually to update all different CF and Columns.
but what about the joins ? is still fast enough when I do the joins manually ?
manually I mean run a query – get list of data extract values considered as keys for other CF and run a query on another CF (and sometimes on a third CF) ?

– Yonatan

By: Yonatan

Yonatan — Thu, 25 Feb 2010 15:06:32 +0000

Any answer’s for Miguel Cobá’s question ?

By: xxb

xxb — Wed, 24 Feb 2010 12:09:27 +0000

nice article, this helps me smooth out confusions of cassandra data model.

By: NOSQL introduction | Thomas Koeppen Blog

NOSQL introduction | Thomas Koeppen Blog — Mon, 08 Feb 2010 22:09:20 +0000

[...] WTF is a SuperColumn? An Intro to the Cassandra Data Model Sep. 2009 Author: [...]

By: josh

josh — Fri, 05 Feb 2010 17:13:56 +0000

This is really helpful, thanks for spending the time.

By: buckmin

buckmin — Wed, 27 Jan 2010 14:01:40 +0000

Nice post.

I have experimented with some schemaless databases i think i am starting to learn how to think non-RDBMS. Still, there are some concepts that bug me. Like ones Miguel mentioned and more importantly just a little more complex queries like ‘posts with guitar and movie tag’ or ‘posts with guitar tag but not with movie tag’. I read in somewhere that we need to intersect or union the ‘grab last posts with guitar tag’ and ‘grab last posts with movie tag’ results manually. Coming from a RDBMS realm this seems too odd for me. Is there a more elegant way of doing this, apart from creating another CF for posts with two tags? I don’t know, how far can we go denormalizing everything?

By: Ive Comu

Ive Comu — Wed, 06 Jan 2010 17:27:23 +0000

Arin,

Nice explanation – thank you for your time.

If you could also include some code (in java / php …) to do gets/ multigets / puts , that would clarify the “group’em” part even more.

Thanks again