So far, we've used the INSERT statements to add new rows to our tables, and the UPDATE statements to update information in existing rows. As it turns out, both INSERT and UPDATE statements can modify existing rows and can create new rows. At their core, we can most accurately think of the INSERT and UPDATE statements as providing different syntax for the same underlying operation, an upsert.

This is quite astonishing for those of us who are used to SQL, in which the INSERT and UPDATE statements are entirely distinct. While there are some situations in which upsert behavior is quite handy, it can also be a stumbling block, especially for developers who are new to Cassandra. Fortunately, Cassandra offers us ways to ensure that our write operations behave the way we intend; we'll explore these techniques after taking a closer look at upserts.