Let's say bob decides he doesn't want his location on his profile. We'd like to get the location column back to the state it was in earlier, which appears as null in cqlsh but simply means "no data here".
As we discussed previously, Cassandra doesn't have the concept of NULL columns in the SQL sense. Concretely, the following statement is not possible in Cassandra:
SELECT * FROM "users" WHERE "location" IS NULL;
Relational databases typically store a separate bit in each column to indicate whether that column contains a NULL value. In Cassandra, on the other hand, all data columns are optional, and only the columns with a value are represented in storage. We can visualize Cassandra rows as maps of key-value pairs, with some keys possibly missing; relational database rows are more like fixed size lists of columns, where some columns may have a NULL value. The following diagram shows how we might visualize dave's row in Cassandra, and...