Rails anti-patterns: how not to design your database

Get Started. It's Free
or sign up with your email address
Rocket clouds
Rails anti-patterns: how not to design your database by Mind Map: Rails anti-patterns: how not to design your database

1. [email protected]

1.1. 60,000,000 rows scanned/sec

2. Anti-Patterns

2.1. Naive trees

2.1.1. Alternatives

2.1.1.1. Closure Trees

2.1.1.1.1. gem 'closure_tree'

2.1.1.2. Materialized Path

2.1.1.2.1. Microsoft uses it in their SQL

2.1.1.2.2. gem 'ancestry'

2.1.1.2.3. Potential for data mistakes

2.1.1.2.4. Fast

2.1.2. hardest to query

2.1.3. joineception

2.1.3.1. joining on every level of tree

2.1.4. N+1 queries

2.1.5. Recursion

2.1.5.1. not very supported in ORM, hard to performance tune

2.1.5.2. not ideal as we have to write recursion to parse the data for displaying them

2.1.6. Simple

2.1.7. Referential Integrity

2.1.8. Efficient on disk space

2.1.9. Use when don't need descendentes

2.2. Polymorphic Relationships

2.2.1. Built-in in Rails

2.2.2. Easy to add more

2.2.3. No foreign keys

2.2.4. No cascading delete

2.2.5. Alternatives

2.2.5.1. Exclusive Arcs

2.2.5.1.1. Foreign key constraints

2.2.5.1.2. Cascading delete

2.2.5.1.3. Possible mistakes

2.2.5.1.4. More migrations

2.2.5.1.5. Many indexes

2.2.5.2. Base Parent Table

2.2.5.2.1. Use when lot of shared fields