Data Views for Data Architecture

Use persistence mode Data Views to visualise conceptual, logical, and physical data models from a single central model. Assign database profiles from 13 vendor templates (including cloud and NoSQL databases) or create custom profiles. Click any node to open the Entity Overlay with the full property table.

January 23, 2026

Data Views in persistence mode replace traditional static ERD diagrams with an interactive, model-derived approach. Entities appear as nodes on a force-directed graph canvas with edges inferred from foreign key relationships. Click any node to open the Entity Overlay, a floating property table that shows key roles, field names, concrete database types, nullable flags, unique constraints, defaults, FK references, and comments. All entity data comes from the central model (model.neoarc); the view stores only layout positions and profile configuration.

Conceptual, Logical, and Physical Data Views

The same model entities produce different levels of detail depending on how the Data View is configured.

Multiple Data Views from One Model

A project typically maintains several Data Views derived from the same model, each targeting a different database vendor, a different subset of entities, or a different audience.

View	Mode	Profile	Entities	Purpose
E-Commerce SQL Server	Persistence	SQL Server (PascalCase)	Customer, Order, OrderItem, Product, Category	Physical schema for SQL Server deployment
E-Commerce PostgreSQL	Persistence	PostgreSQL (snake_case)	Customer, Order, OrderItem, Product, Category	Physical schema for PostgreSQL deployment
Payment Domain	Persistence	SQL Server (PascalCase)	Payment, Invoice, Refund, Transaction	Focused view of the payment subdomain
Product Search	Search	Azure Cognitive Search	Customer, Product, Category	Search index configuration and drift detection
Logical Overview	Persistence	(Abstract Types)	All entities	Database-agnostic logical model for design review

The Entity Overlay

Click any node on the graph canvas (or press Alt+Enter) to open the Entity Overlay. The overlay is a floating panel that shows the full property table for that entity, with columns determined by the view mode.

Column	Persistence Mode	Search Mode
Key Role	PK, FK, NK, SK, CP indicators	Key field indicator
Field Name	Converted via profile naming convention	Search field name
Type	Concrete database type (e.g. NVARCHAR(100))	Search field type (e.g. Edm.String)
Nullable	Yes/No	n/a
Unique	Yes/No	n/a
Default	Default value expression	n/a
FK Reference	Target entity and property	n/a
Behaviours	n/a	Searchable, filterable, sortable, facetable, retrievable

Relationship pills on either side of the overlay let you navigate to connected entities. Pin an overlay to dock it below the canvas, then click another node to compare two entities side by side.

Example: E-Commerce Data Model with Two Profiles

Consider an e-commerce model with Customer, Order, OrderItem, Product, and Category entities. Creating two persistence mode Data Views with different database profiles produces vendor-specific physical representations from the same source data. The Entity Overlay for each view shows the resolved types.

Property	Abstract Type	SQL Server	PostgreSQL
customerId	uuid	UNIQUEIDENTIFIER	UUID
firstName	string (maxLength: 100)	NVARCHAR(100)	VARCHAR(100)
email	string (maxLength: 255)	NVARCHAR(255)	VARCHAR(255)
isActive	boolean	BIT	BOOLEAN
createdAt	datetime	DATETIME2	TIMESTAMP WITH TIME ZONE
metadata	json	NVARCHAR(MAX)	JSONB
orderTotal	decimal	DECIMAL(10,2)	NUMERIC(10,2)

Graph Canvas Features

The Data View graph canvas provides interactive tools for exploring the data model.

Feature	Description
Force-directed layout	D3 physics simulation positions nodes automatically based on relationships
Spacing slider	Adjust graph density in real time via the bottom-right slider
Node search (Ctrl+F)	Fuzzy search with ranked results - exact, prefix, then fuzzy matches
Semantic zoom	Node detail increases as you zoom in; zoomed-out shows labels only
Group hulls	Coloured boundaries around entity groups defined in the model
Minimap	Overview panel for navigating large graphs
Pinning	Fix important nodes in place while the simulation runs around them

Fidelity Considerations

Not all abstract types map equally across vendors. When switching a view from one database profile to another, the system checks for fidelity loss and shows a confirmation dialog listing any warnings.

Abstract Type	Fidelity Concern	Affected Vendors
uuid	Stored as text (CHAR(36)) rather than native binary	MySQL
json	Stored as text (NVARCHAR(MAX)) without native validation	SQL Server
boolean	Emulated with TINYINT(1)	MySQL
enum	No native enum type, requires CHECK constraint	SQL Server, Oracle

Naming Conventions Across Profiles

Each database profile includes a naming convention that controls how entity and property names are displayed in the Data View. The model stores names in their original format; conversion is applied at display time only.

Model Name	SQL Server (PascalCase)	PostgreSQL (snake_case)	Oracle (SCREAMING_SNAKE_CASE)	MongoDB (camelCase)
CustomerOrder	CustomerOrder	customer_order	CUSTOMER_ORDER	customerOrder
firstName	FirstName	first_name	FIRST_NAME	firstName
orderTotal	OrderTotal	order_total	ORDER_TOTAL	orderTotal

Using Groups to Organise Views

When the model defines groups (subdomains), the entity picker provides group filtering. This simplifies creating views that focus on a specific domain area. For example, an e-commerce model might have groups for Core Commerce, Product Catalogue, and Logistics, each with a corresponding focused Data View.

SQL DDL Import

The SQL DDL import reads existing database scripts and populates the central model with entities, properties, key roles, and relationships. Paste or upload DDL text from 6 dialects (PostgreSQL, MySQL, SQL Server, Snowflake, Oracle, and DBML). From the imported model, Data Views can be created to visualise the schema interactively.