Patrik Nordwall

@krasserm @zapletal-martin Now, when Cassandra 3.0 is released with the new Materialized Views, do you still see any reason for not using them? The EventsByTag query is important for the...

so you are saying that MVs in Cassandra 3.0 is not production ready?

Thanks for the information. Anyway, Akka Persistence Query is also experimental, and I doubt that we will be able to implement manual materialized views in a better way than what...

That is a good point. Another thing I'm uncertain about is what we can use as offset column. I was thinking of a `timeuuid`, but that might be too weak?

Interesting questions, indeed. They discuss consistency in length in [CASSANDRA-6477](https://issues.apache.org/jira/browse/CASSANDRA-6477), but I'm not sure what the conclusion or implementation is. I assume that MV is updated in the same order...

I have also put together a prototype of EventsByTag query using materialized view. I see no blockers for the needs I have in my current project. > No causal consistency...

Let me clarify with an example. We have persistent actors A and B. 1. persist event a1 and a2 in A 2. send a comand from A to B 3....

Then I'm sorry that I completely misunderstood what you were refering to with casaul consistency. I thought you meant across different persistentIds. I'm aware about the async nature of MV...

I'll create the branch tomorrow morning and open the first pull request.

I have implemented `eventsByTag` using materialized views. It's ready as far as I can see. Please review. Let's discuss in the PR. https://github.com/krasserm/akka-persistence-cassandra/pull/128