Denormalizing Data for Maximum Performance - Learning Apache Cassandra - page 118

A normalized approach
Before we proceed down the path of denormalization, let's first try an approach that re-
quires adding no new tables to our schema. Our goal is to display to a user all the status up-
dates of all the users they follow, the most recent first. So, the simplest approach would be
to simply look up the followed users, and then retrieve their status updates.
If you've been following along with the code examples so far, you should currently have a
single follow relationship in the database, such as alice follows carol . To make things
more interesting, let's also have alice follow dave :
INSERT INTO "user_outbound_follows"
("follower_username", "followed_username")
VALUES ('alice', 'dave');
INSERT INTO "user_inbound_follows"
("followed_username", "follower_username")
VALUES ('dave', 'alice');
Now that alice is following a couple of users, we can build a meaningful home timeline
for her. Let's have carol and dave write some updates:
INSERT INTO "user_status_updates" ("username", "id", "body")
VALUES ('dave', NOW(), 'dave update one');
INSERT INTO "user_status_updates" ("username", "id", "body")
VALUES ('carol', NOW(), 'carol update one');
INSERT INTO "user_status_updates" ("username", "id", "body")
VALUES ('dave', NOW(), 'dave update two');
INSERT INTO "user_status_updates" ("username", "id", "body")
VALUES ('carol', NOW(), 'carol update two');