It is also possible to declare a table that has both a composite partition key as well as multiple clustering columns. Suppose we wish to model a status update replies table to be clustered by the reply date first and then the reply time; we would have a table like this:
CREATE TABLE "status_update_replies" ( "status_update_username" text, "status_update_id" timeuuid, "status_date" date, "status_time" time, "id" timeuuid, "author_username" text, "body" text, PRIMARY KEY ( ("status_update_username", "status_update_id"), "status_date", "status_time", "id" ) );
In the preceding table, the partition key is a combination of status_update_username and status_update_id.
The replies ...
