Now let's take a look at the counterpoint problem to deletes, when data for a given key grows without bounds. This is a surprisingly easy situation to get yourself into, especially if you do not understand how Cassandra stores your data on disk. Perhaps the best antidote to unbounded row growth is to read and understand the previous chapter, which offers the foundational knowledge to help you avoid this scenario.

To clarify, this section is not a warning against unbounded growth of your dataset in general. We have established that Cassandra scales linearly, so you can continue to add data as long as you have capacity in your cluster. Instead, we are referring to a model where a given partition key continues to accumulate columns with no end in sight. We briefly touched on this in Chapter 7, Modeling for High Availability, but the topic deserves a full treatment.

We can imagine a typical scenario using the sensor_readings data model described in the previous chapter. Here...