NoSQL Data Models

NoSQL Data Models

1. Application specific

2. Aggregates

   • Combine simple data types to complex
   • Soft schema
     • Constraints enforced by application layer
     • 1-M relationships less

3. Modelling

   • Enumerable keys
     • Grouped by value
   • Atomic attributes
     • Limited transaction supprot
     • Storage of entity and update
   • Dimensional Reduction
     • Multi-dimension data to key-value
       • E.g Geohashing
   • Index Table
     • Indexes are data stores
     • Table with keys
   • Composite Key Index
     • Multiple attributes
     • Multi-dimension indexing
     • Grouped for aggregation

4. Inverted Search

   • Data processing
   • Index to search criteria
   • Aggregate data

5. Hierarchy Techniques

   • Tree aggregation
     • Single record or doc
     • Single access efficience
     • Updates are inneficient
   • Adjacency List
     • Nodes are independent records
     • Search via indentifiers
     • Inefficient for entire subtree
   • Materialised Paths
     • Node has identifiers of parents or children
     • Stored as IDs
   • Nested sets
     • Stored in array
     • Efficient for immutable data
     • Inset & updates costly
   • Batch graph
     • Good for neighbour search
     • Inefficient