Skip to content
CSA Loom — the Microsoft Fabric experience for Azure tenants where Fabric isn't yet available: lakehouses, warehouses, notebooks, semantic models, Activator rules, Data Agents, across Commercial, GCC, GCC-High, and DoD IL5

Tutorial: Synthetic data editor

CSA Loom synthetic-data editor — generate realistic-looking but entirely fabricated rows from per-column strategies, preview them, then write a real Delta table through Databricks SQL. PII-classified source columns are mapped to synthetic strategies — fake names and emails, never real data. Azure-native — no Microsoft Fabric.

What it is

Development, demos, load tests, and training courses all need data that looks real without being real. This item lets you declare a schema — either by hand or seeded from a data contract — assign each column a generation strategy, preview a sample, and then materialize a full table (up to 200,000 rows) into Unity Catalog as Delta.

Two tabs: Design and Runs.

When to use it

  • You need a populated table in a lower environment without copying production.
  • You want the shape of a data contract's schema, with none of its PII.
  • You need reproducible data — the same seed produces the same rows.

Step-by-step in Loom

  1. Create the item. + New item → Synthetic data, then Create synthetic data generator.
  2. Choose a source schema. In Source schema pick either:
  3. Define columns — start from an empty grid, or
  4. From a data contract — pick a contract in this workspace; its columns seed the grid and strategies are inferred, with PII-classified columns mapped to synthetic strategies automatically.
  5. Set a strategy per column. In Columns & generation strategy, each row has a name, a type, a strategy, its options, and a null-rate percentage:
Family Strategies Options you fill
Keys Sequence (auto-increment), UUID start at
Numeric Integer (uniform range), Decimal (uniform range), Number (normal distribution) min/max, decimals, mean/stddev
Temporal Date (range), Timestamp (range) start / end
Discrete Boolean, Categorical (from values), Constant comma-separated values, value
Synthetic identity Full name, First name, Last name, Email, Phone, Street address —
Synthetic other Company, City, Country, Redacted (mask token) —

Add column and the per-row delete button manage the grid. 4. Set volume and reproducibility. Rows to generate (up to 200,000) and Seed — the same seed reproduces the same rows exactly. 5. Choose the write target. In Write target, cascading dropdowns of your real Databricks estate: SQL warehouse → Catalog → Schema → Staging volume (the UC volume used to stage the rows), plus a Table name for the new Delta table. 6. Preview. Preview sample generates up to 10 real rows through the backend and renders them in a table. Preview works even with no Databricks backend configured — only the write needs one. 7. Generate. Generate table saves any pending edits, generates the full row set, stages it to the volume, and creates the real Delta table through Databricks SQL. 8. Check the record. Runs lists every generation: started, target, requested rows, rows written, and status (succeeded / partial / failed) with the error when one occurred.

The Azure backend it rides on

  • Write path: Databricks SQL over Unity Catalog — rows are staged to a UC volume and registered as a real Delta table (the createUcTableFromFile path).
  • Generation + preview: the server-side generator (POST …/preview, POST …/generate).
  • Catalog browsing: GET …/catalog lists the real warehouses, catalogs, schemas, and volumes you can reach.
  • Persistence: the item's own state holds the column specs, target, and run history.

Honest gates

Condition What you see Exact remediation
Databricks not configured Warning MessageBar "Databricks not configured" naming the exact env var, and stating plainly that preview still works Set the named variable (plus a SQL warehouse) on the loom-console env
Target incomplete Generate table disabled with "Add columns and pick a warehouse, catalog, schema, volume and table to generate." Complete the cascade
No data contracts in the workspace The From a data contract radio is disabled with a hint Author a data-contract item with a schema first
Generation partially fails The run is recorded as partial with rows written and the error Read the error, adjust the spec, re-run

PII posture

Strategies flagged as synthesizing PII (full name, first / last name, email, phone, street address) produce fabricated values. When columns are seeded from a data contract, PII-classified columns are mapped to those synthetic strategies automatically — no real data is ever read or copied.

No Fabric required

Databricks SQL + Unity Catalog + Delta. No Fabric capacity, workspace, OneLake path, or Power BI workspace is used on any path.

Learn more