explainer · published
Conceptual, logical and physical data models: which tools cover each level
A short refresher on the three model levels, then a table of which data modeling tools say they support each one, from the vendors' own pages.
What are the three levels?
The three levels describe the same data at increasing detail.
- Conceptual: the business things (customer, order, product) and how they relate. No columns or data types. The audience is people who know the business.
- Logical: entities with attributes, keys and relationships, still independent of any database product.
- Physical: tables and columns for a specific database, with exact data types, indexes, constraints and naming rules. This is what gets turned into DDL.
Vendors use slightly different words. Oracle, for example, describes logical, relational, physical and multidimensional models in SQL Developer Data Modeler. The idea is the same.
Do you need all three?
Not always. If you mostly document databases that already exist, physical-only tools like dbdiagram.io, DrawSQL or DBeaver are enough.
You need the upper levels when the model has to outlive a particular database, when business people review it, or when one logical model feeds several physical targets. That’s typical of warehouses and enterprise architecture.
Which tools say they support which level?
The table uses only what each vendor states on the pages we read. A “not verified” cell means the page didn’t say, not that the tool can’t do it.
| tool | conceptual | logical | physical | DDL out |
|---|---|---|---|---|
| ChartDB | no | not verified | yes | yes |
| dbdiagram.io | no | not verified | yes | yes |
| DBeaver | no | not verified | yes | not verified |
| DbSchema | yes | yes | yes | yes |
| draw.io (diagrams.net) | yes | not verified | no | no |
| drawDB | no | not verified | yes | yes |
| DrawSQL | no | not verified | yes | yes |
| Ellie | yes | yes | yes | not verified |
| ER/Studio | yes | yes | yes | yes |
| Hackolade Studio | not verified | not verified | not verified | not verified |
| Lucidchart | yes | not verified | not verified | not verified |
| Luna Modeler | yes | yes | yes | yes |
| Mermaid (erDiagram) | yes | yes | no | no |
| Moon Modeler | not verified | not verified | yes | yes |
| MySQL Workbench | no | not verified | yes | yes |
| Navicat Data Modeler | yes | yes | yes | yes |
| Oracle SQL Developer Data Modeler | yes | yes | yes | yes |
| pgModeler | no | not verified | yes | yes |
| QuickDBD | no | not verified | yes | yes |
| Redgate Data Modeler | not verified | yes | yes | yes |
| SqlDBM | yes | yes | yes | yes |
| Visual Paradigm | yes | yes | yes | yes |
Which editions do the levels come in?
Watch for levels sold separately. DbSchema’s purchase page says its Architect edition “adds logical and conceptual design”. Redgate Data Modeler lists “Physical & Logical design” only with its higher-priced tier. Navicat Data Modeler describes converting “a conceptual model to logical, then to physical” in one workspace.
Check the specific edition before buying, not just the product name.
Sources
Update history
- Published.
Tools in this guide
-
- id
- er-studio
- price
- from $2,687 subscription
- deploy
- not verified
- license
- proprietary
Enterprise modeler covering conceptual, logical and physical models, with DDL generation, reverse engineering and model compare.
-
- id
- navicat-data-modeler
- price
- from $749 one-time
- deploy
- Windows, macOS, Linux
- license
- proprietary
Cross-platform desktop modeler with conceptual, logical and physical models plus relational, dimensional and Data Vault 2.0 methods.
-
- id
- luna-modeler
- price
- from $99 one-time
- deploy
- Windows, macOS, Linux
- license
- proprietary
Relational sibling of Moon Modeler: conceptual, logical and physical models, DDL generation, and compare-and-sync against a live database.
-
- id
- ellie
- price
- from €49/mo
- deploy
- Web
- license
- proprietary
Browser-based business and semantic modeling tool: conceptual through physical models plus a business glossary, with AI-assisted modeling.
Related guides
-
- slug
- how-to-choose-a-data-modeling-tool
- type
- how-to
- updated
Seven questions, asked in order, that cut a long list of data modeling and ERD tools down to two or three you should actually trial.
-
- slug
- diagram-as-code-dbml-mermaid-sql
- type
- worked example
- updated
The same two-table schema written three ways, with the SQL that DBML actually generates, and what each format is good and bad at.