To gain read performance, data model in Cassandra are denormalized to certain extend.
As denormalized produces duplicated record, how to avoid data anomalies?
To gain read performance, data model in Cassandra are denormalized to certain extend.
As denormalized produces duplicated record, how to avoid data anomalies?
据我所知,可能有两种方法可以使您的数据在不同的列族(CF)中保持一致:
您想使用哪种方式,这可能取决于您如何对数据进行建模,以及您的应用程序对不同 CF 之间数据一致性的要求有多严格。
jsfeng 的回答在一般原则上很好,但更具体的回答是使用原子批处理:www.datastax.com/dev/blog/atomic-batches-in-cassandra-1-2