Database Reference
In-Depth Information
foo@bar.com | {756716f7... } | {'AMZN':
'2013-06-13 11:42:12-0400',
'GOOG': '2013-06-13
12:51:31-0400'}
DELETE ticker_updates['AMZN']
FROM users
WHERE email = 'foo@bar.com';
email | portfolios | ticker_updates
.............+................+.....................................
foo@bar.com | {756716f7...} | {'GOOG':
'2013-06-13 12:51:31-0400'}
Summary
Data modeling in Cassandra may seem counterintuitive to someone who is used to
a relational database. Certain concessions need to be made to gain performance in
high-volume workloads. The main points to take away from this chapter are that
relational database modeling techniques will almost never apply; model your quer-
ies, not your data; and denormalization and duplication of data are not bad—in
fact, they are recommended. Also, keep in mind that “normalization” of your data,
in any way, is almost never recommended. Collections can be very powerful, but
they may impact performance when it comes to very large data sets.
