ADO'NET 3'0 Entity Data Model

1
(No Transcript)
2
ADO.NET 3.0 Entity Data Model

• Gert E.R. Drapers
• Chief Software Architect
• Visual Studio Team Edition for Database
  Professionals
• Microsoft Corporation
• http//blogs.msdn.com/gertd

3
Agenda

• ADO.NET Entity Framework Introduction
• Application Scenarios
• Direct Mapping
• Flexible Mapping Scenarios
• Extensibility
• Usage Best Practices
• Summary

4
The Entity Framework
The Entity Data Model
ADO.NET 2.0
Dataset
ADO.NET Providers
Language Integrated Query (LINQ)
ltbookgt lttitle/gt ltauthor/gt ltyear/gt
ltprice/gt lt/bookgt
5
Entity Framework Overview
EDM MetaData
LINQ to Entities
Object Metadata
Object Services
O-C Map
Conceptual Schema
C-S Map
ADO.NET Data Provider
Store Schema
ADO.NET Data Provider
ADO.NET Data Providert
6
EDMGEN.EXE

• Generates the model files
• Model.ssdl (Store Schema Definition Language)
• Describes tables and columns which map to
  Entities and Relationships
• Model.csdl (Conceptual Schema Definition
  Language)
• Describes the Entity Data Model (incl.
  EntitySets, EntityTypes, Associations
  AssociationsSets)
• Model.msl (Mapping Specification Language)
• Describes how the Entity Framework maps between
  the Conceptual Model (CSDL) and the logical
  Storage Schema (SSDL)
• Model.cs or Model.vb
• Partial class implementing the model created

7
Direct Mapping

• Scenario
• Database schema provides appropriate application
  model
• Benefits
• Relationships
• Common Extended Query Language
• Common Queryable Schema
• Common Metadata Services
• Object Services
• Business Logic
• Identity Management
• Change Tracking
• Language Integrated Query (LINQ)

8
Flexible Mapping

• Scenario
• Database schema doesn't match application model
• Benefits
• Richer Application Data Model
• Inheritance
• MN Relationships
• Richer Mapping
• Inheritance options (TPH, TPT, TPC)
• Stored Procedures
• Entities split across tables
• Decouples application model from storage schema
• Multiple application models for same store
• Same application model over different storage
  schemas
• Independent evolution of application
  model/storage schema

9
Extensibility

• Scenario Building a customizable framework
• Benefits
• Flexible run-time mappings
• i.e., Language-specific mappings
• Mapping to "Extensible Schemas"
• EntityViews
• Vertical partitioning of framework and extensions
• Separate mappings
• Derived types
• Extending data classes

10
Best Practices

• Should I use Objects or DataRecords?
• How long should I hold an ObjectContext?
• Which mapping techniques should I use?
• When should I used Stored Procedures or Table
  Valued Functions?
• How should I extend data classes?
• When should I write my own data classes?
• When should I use Custom Serialization?
• Should I use LINQ, Query Builder Methods, or
  eSQL?
• Should I use LINQ to SQL or LINQ to Entities?

11
Objects or DataRecords?

• Objects (IEnumerableltTgt)
• Strong typing
• Business logic
• Updating
• Example Interactive Windows application
• Data Records (IExtendedDataReader)
• Performance for read-only streaming
• Sequential access
• Avoid object construction overhead
• Doesn't require generated classes
• Example Directly serializing results to a Web
  response

12
Holding ObjectContext

• Web scenarios
• Typical pattern Create, Use, dispose
• ObjectContext holds resources
• Connections
• Object References
• Use NoTracking for read-only scenarios
• Windows Forms scenarios
• Hold as long as application interacts with
  results
• ObjectContext mantains state
• Identity resolution
• Change tracking
• New ObjectContext for each new query
• Ensures data is fresh

13
Mapping Techniques

• Database views
• Common access patterns across applications
• Optimization of access paths through Materialized
  Views
• Table/View mappings
• Entity properties map directly to database
  columns
• Columns have same meaning throughout hierarchy
• EntitySQL views in MSL
• Flexible mappings expressed in EntitySQL
• Filter based on conditions, read-only properties,
  non-equality conditions, computed properties,
  multiple extents mapped to same table
• Native SQL views in SSDL
• Mapping on top of native SQL queries
• Store-specific queries
• Expose as "EntitySet" to mapping layer

14
Stored Procedures and TVFs

• Stored Procedures
• Better control over operations on data within the
  database
• Tighter control over operations, permissions
• Provides level of indirection from storage schema
• Use for operations that affect state of the
  database
• Inserts, updates, deletes
• Required for updates against ESQL and Native SQL
  views
• Can return multiple results in one roundtrip
• Tabled Valued Functions
• Better control over access to data within the
  database
• Tighter control over operations, permissions
• Provides level of indirection from storage schema
• Use for server-side functions that return data
• Results of TVF are composable within a query

15
Extending Data Classes

• Derivation
• Adding new persistent fields to a base class
• Derived types
• Customization extensions
• Partial Classes
• Business logic
• Methods
• Calculated values
• Non-persistent fields
• Custom serialization
• "Activity" Classes
• Common operations on data classes

16
Custom Data Classes

• Entity Framework-generated data classes
• Derive from Entity Framework base class
• Change Tracking, RelationshipManagement
• Common Notifications, Property Validation
• Custom data classes
• Derive from Entity Framework base class
• Generate common patterns
• Insert common intermediate class
• Derive from your own base class
• Implement IEntity/IEntityWithRelationships
• Expose EntityKey
• Notify ObjectContext when data changes
• RelationshipManager property to do relationship
  fixup
• Add attributes for persistent fields
• How we get O/OC metadata today
• Investigating "Persistence Ignorance" scenarios

17
Custom Serialization

• Default serialization
• Serialize streams of Entities
• Doesn't navigate relationships in order to avoid
  cycles
• Custom serialization
• Serialize Graphs
• Manage relationship serialization to avoid cycles
• Serialize to different shapes
• "Entity Projections"

18
LINQ, Query Builder, or eSQL?

• LINQ
• Strongly typed queries
• Compile-time checking
• Embedded in your application
• LINQ-Mapped Canonical Functions
• ObjectQueryltTgt builder methods
• Queries built from user input
• Iterative query building
• Store functions
• eSQL
• Dynamic/AdHoc queries
• Persistent queries
• Store functions

19
LINQ to SQL or LINQ to Entities?

• LINQ to SQL
• Emphasis on rapid application development
• Direct mapping
• Single type hierarchy to single table/view/stored
  proc/TVF
• Simple renaming
• Derive class for update logic/Stored Procedures
• Direct mapping to Microsoft SQL Server family of
  databases
• Microsoft Visual Studio 2008 RTM

20
LINQ to SQL or LINQ to Entities?

• LINQ to Entities
• Focus on enterprise-grade data scenarios
• Flexible Mapping to Relational data
• Mapping a single class across multiple
  tables/views
• Mapping to different types of inheritance
• Directly modeling Many to Many relationships
• Mapping to an arbitrary query against the store
• Common model across products
• Declarative stored procedures
• Access to Microsoft SQL Server family and other
  databases through ADO.NET provider model
• Microsoft Visual Studio 2008 update

21
The LINQ Project
.NET Language Integrated Query
22
Summary

• Entity Data Model provides an application-oriented
  data model
• Entities and relationships
• Strong typing, polymorphism
• Client Views provide flexible mapping to your
  storage schema
• Entities split between tables, ManyMany
  relationships, mapping to store queries, etc.
• Entity Framework lets you opt-in according to
  your requirements
• DataReaders/Objects
• eSQL/QueryBuilder/LINQ
• Generated/Custom Data Classes

23
? Questions ?

• http//blogs.msdn.com/gertd