Data manipulation operations in Apache Cassandra™ include inserts, updates and deletes, and have respective CQL statements
DELETE. They are used to add, change, and remove rows and column values in Cassandra tables.
While their general semantics is easy to understand and their syntax is similar to the respective operations in relational databases
and SQL, there are also important differences.
All three operations are write operations. Inserts write new data into tables. Similarly, updates write new data into tables. Even deletes write markers, called tombstones, to signify that data is now removed. Since there is no read before a write, the operations do their writes regardless of whether a row with the same primary key already exists in a table or not. Consider the following three situations. First, when inserting a new row into a table and the table already has a row with the same primary key, the existing row will effectively be updated. In other words, the insert operation will upsert data into the existing row. Second, when updating a row that does not exist in a table, the row will be inserted into the table. In other words, the update operation will upsert the new row into the table. You can think about an upsert as an insert that causes an update or an update that causes an insert. Finally, when deleting a row that does not exist in a table, a tombstone will simply be written.
It is also worth mentioning, that Cassandra does support more expensive conditional inserts, updates and deletes that implement the read-before-write pattern. We will see examples of those, too.