cassandra-basics

Basic example of data modelling and implementation for Cassandra, including:

• SASI (to allow complex queries, such as contains, suffix, etc.)
• Materialized views (not recommended since it increases write time)
• Range queries
• Table with TTL
• Batch statements (to ensure consistence between tables)

Business data model

A vessel has several attributes:

• UUID: unique technical identifier
• Name: display label, not necessarily unique
• Category: vessel category (cargo, windsurf, etc.)
• Visibility: either "all centers", or "creation center only"
• Creation center: ID of the center that created the vessel instance
• Last departure port: ID of the port that the vessel last departed from (only applicable to category cargo)
• Last departure time: time that the vessel departed from last port (only applicable to category cargo)

Queries

• READ: Get list of all vessels that are visible to site
• READ: Get vessel from selected UUID
• READ: Find vessels by name fragment (among vessels that are visible to site)
• READ: Find vessels by category (among vessels that are visible to site)
• READ: Get list of vessels that departed recently from a selected port
• WRITE: Create vessel
• WRITE: Update vessel departure info
• WRITE: Change vessel visibility
• WRITE: Delete vessel

Physical data model

Namespace: vessel

Get list of all vessels that are visible to site

• Table name: vessels
• Partitioning key(s): "_ALL" or creation center ID
• Clustering key(s): UUID

A secondary index (SASI) is created against name attribute of vessels table, in order to be able to find rows that "contain" name fragments.

Materialized view vessels_by_category is created to request vessels against category attribute of vessels table.

Get vessel from selected UUID

• Table name: vessels_by_uuid
• Partitioning key(s): UUID
• Clustering key(s): (none)

Get list of vessels that departed recently from a selected port

• Table name: vessels_by_departure_port
• Partitioning key(s): last departure port
• Clustering key(s): last departure time, UUID

Only vessels with last departure info available are inserted in this table.

Last departure time clustering column is used to order results, as well as to perform range queries, so that obsolete data is filtered out. A default TTL is defined at the definition of the table so that obsolete content does not stay in the table indefinitely.

How to run

• Run tests from IDE / Maven build