Skip to content
Tutorials Center
My learning

Topics

Lesson 16 of 38 ยท Data and Dataverse

Dataverse Tables and Columns

Learn Dataverse basics for Power Apps: standard, custom and activity tables, primary columns, column data types, calculated and rollup columns, keys and views.

Microsoft Dataverse is the data platform behind the Power Platform. It stores data in tables made of columns, adds relationships, security and business logic, and exposes everything to Power Apps, Power Automate and Power BI. This lesson covers how to design tables and pick column types.

Dataverse is a cloud service, so it cannot run here. Steps are described in words and tables; the stand-ins model the data design ideas in JavaScript.

Tables

A table (formerly called an entity) holds rows of one kind of thing. Types:

  • Standard tables: ship with Dataverse, such as Account, Contact and User. They support common business scenarios, and many Microsoft apps rely on them.
  • Custom tables: tables you create, such as Project or Inspection.
  • Activity tables: for things like emails, appointments and tasks.
  • Virtual tables: show data from an external system without copying it.
  • Elastic tables: for very large, fast-growing data like telemetry (a specialized option).

To create one: maker portal, Tables, New table, choose a name, and optionally customize options such as attachments and audit. Every table gets a primary key (a GUID) and a primary column (the human-readable name shown in lookups).

The table has a display name (Inspection), a logical name with a publisher prefix (cr123_inspection), and an entity set name used by the API. Your solution publisher supplies the prefix.

Columns and data types

Common column types:

Type Notes
Text (single line) Up to 4,000 characters; formats like email and URL
Multiline text Long notes
Whole number, Decimal, Float, Currency Choose precision carefully
Date and time Choose date only or date and time; time zone behavior matters
Yes/No A two-option choice
Choice and Choices List of options (see Relationships and choices)
Lookup Link to a row in another table
File and Image Binary data
Autonumber Generated sequences, such as INS-0001
Formula, Calculated, Rollup Derived values

Each column has a display name, logical name, requirement level (Optional, Business recommended, Business required), and optional searchable/auditing settings.

Calculated, rollup and formula columns

  • Calculated column: computed on read from other columns of the same row (and parent lookups). Example: FullName = FirstName + " " + LastName.
  • Rollup column: aggregates related child rows, such as the sum of order totals per customer. Recalculated on a schedule (about hourly) or on demand.
  • Formula column: written in Power Fx and calculated on read, a modern alternative to calculated columns.

These move logic to the server so apps and flows all see the same value.

Keys and uniqueness

Dataverse rows are identified by a GUID. For uniqueness on business data, define an alternate key, for example on EmployeeNumber, which also supports upsert operations from integrations.

Views, forms and security

A table also owns:

  • Views: saved filters and column layouts (for example "Active inspections").
  • Forms: layouts for editing rows (used by model-driven apps).
  • Business rules: simple logic on forms.
  • Security: table-level privileges by security role (Create, Read, Write, Delete, Append, Assign, Share) and access scope (user, business unit, organization). Field-level security can protect single columns.

Modeling exercise

Suppose an equipment inspection app needs a table Inspection with the columns Title (primary), Site (lookup), InspectedOn (date), Status (choice) and Score (whole number). The stand-in defines that schema as data and checks a row against it, the way Dataverse enforces types and required columns.

JS Stand-in: schema and row validation

Press Run (or Ctrl + Enter) to see the console output here.

No output.

A rollup example

The next stand-in computes a rollup, the sum and count of child rows per parent, the way a rollup column would.

JS Stand-in: a rollup column (sum and count of children)

Press Run (or Ctrl + Enter) to see the console output here.

No output.

Design guidelines

  • Reuse standard tables (Account, Contact) when they fit instead of inventing near-duplicates.
  • Use prefix and names consistently; display names are for users, logical names for code.
  • Choose column types early; changing type later can be impossible or lossy.
  • Prefer choices over free text for fixed lists.
  • Keep one concept per table; link them with lookups.
  • Avoid storing derived data you can compute with formula or rollup columns.

Common mistakes

  • Using text columns for dates or numbers.
  • Treating Dataverse like a spreadsheet with dozens of repetitive columns (Item1, Item2) instead of a child table.
  • Ignoring security roles until the app is finished.
  • Forgetting that solutions carry the publisher prefix; use one publisher for a project.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. What identifies each row in a Dataverse table internally?

    Every row has a GUID key; the primary column is the readable name.
  2. 2. Which column type sums values from related child rows?

    Rollup columns aggregate related rows, for example the sum of order totals.
  3. 3. What is an alternate key used for?

    Alternate keys enforce uniqueness on business columns and help integrations.
  4. 4. Where does the logical name prefix such as cr123_ come from?

    Each publisher defines a customization prefix applied to its components.

Summary

  • Dataverse stores data in tables with typed columns, keys, views, forms and security.
  • Table kinds: standard, custom, activity, virtual and elastic.
  • Calculated, rollup and formula columns put derived logic on the server.
  • Design with reuse, correct types, choices and separate tables per concept.

Next: Filter, search, sort and Patch