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 02 — First lakehouse + Delta tables

Upload sample data into a lakehouse, query it with Synapse Serverless, and write a Silver Delta table from a notebook. 30 minutes.

Prerequisites

  • Workspace from Tutorial 01
  • Console open on the workspace home

How navigation works

Loom uses top-level left-nav surfaces (Home, Workspaces, Browse, OneLake catalog, Unified catalog, Lineage, Data agents, Monitor, …) plus a flat item tree inside each workspace. Item types are not per-workspace "panes" — you create them with the workspace's New item button, and each one opens as an editor at /items/<type>/<id>.

Steps

1. Open your workspace

Left nav → Workspaces → click your workspace row. You land on /workspaces/<id> with a flat item tree (folders first, then items, sorted alphabetically). It's empty for a new workspace.

2. Create a lakehouse

Workspace header → New item → category Data Engineering → Lakehouse. Give it a name and click Create. You land on the Lakehouse editor at /items/lakehouse/<id>.

3. Pick a container

The editor's left panel shows the ADLS Gen2 container tree (loaded from /api/lakehouse/containers). Pick a container — this is the Azure-native Delta store backing the lakehouse. No Fabric/OneLake workspace is required.

4. Upload sample data

Toolbar → Upload file. Choose a CSV or Parquet sample (e.g. the NOAA daily weather sample from examples/fiab-quickstart/data/noaa-daily-2025-01.csv). The file appears in the tree and a success toast shows Uploaded <name> at HH:MM:SS.

5. Preview the file

Click the uploaded file. The editor auto-populates the SQL tab with a Synapse Serverless OPENROWSET template for that file. Switch to the Preview tab to render the first 100 rows (served by /api/lakehouse/preview).

6. Query it with Serverless SQL

Switch to the SQL tab. Adjust the generated OPENROWSET query if needed and click Run (calls POST /api/items/lakehouse/<id>/query, which executes against Synapse Serverless). Results render in the grid.

7. Open the file in a notebook

Right-click the uploaded file → Open in notebook. This opens a new Notebook editor (/items/notebook/new?lakehouse=…&path=…) prefilled with Spark code that loads the file and writes a Delta table. A representative Bronze cell:

df = spark.read.csv(
  "abfss://<container>@<storage-account>.dfs.core.windows.net/Files/noaa-daily-2025-01.csv",
  header=True,
  inferSchema=True,
)

df.write.format("delta").mode("overwrite").saveAsTable("noaa_bronze_daily")
print(f"Bronze table created with {df.count()} rows")

Run the cell (the notebook executes against the compute backend via POST /api/items/notebook/<id>/run-cell). The Delta table lands in the lakehouse Tables/ prefix.

8. Verify the Bronze table

Return to the Lakehouse editor and open the Tables tab (it lists the Tables/ prefix via /api/lakehouse/paths). noaa_bronze_daily appears.

9. Transform Bronze → Silver

Back in the notebook, add a second cell to clean and partition the data:

from pyspark.sql.functions import col, to_date

silver_df = (
  spark.table("noaa_bronze_daily")
  .withColumn("date", to_date("date_str"))
  .withColumn("temperature_c", (col("temp_f") - 32) * 5 / 9)
  .filter(col("temp_f").isNotNull())
)

silver_df.write.format("delta") \
  .mode("overwrite") \
  .partitionBy("date") \
  .saveAsTable("noaa_silver_daily")

print(f"Silver table: {silver_df.count()} rows, partitioned by date")

Run the cell. The Silver table appears in the lakehouse Tables tab.

10. Aggregate over Silver

In the Lakehouse SQL tab, OPENROWSET (or SELECT against the registered table) the Silver data for a monthly rollup:

SELECT
  YEAR(date) AS year,
  MONTH(date) AS month,
  COUNT(*) AS days_recorded,
  AVG(temperature_c) AS avg_temp_c
FROM noaa_silver_daily
GROUP BY YEAR(date), MONTH(date)
ORDER BY year, month

Click Run. Results render in the grid. There is no separate "Warehouse" surface to switch to — Serverless SQL is built into the lakehouse editor.

11. Catalog the tables

Left nav → Unified catalog. Find both tables. Open noaa_silver_daily and add a description, tags (e.g. domain: weather), and column descriptions. The catalog writes to UC (Commercial) or Purview (Gov) — same UX.

What's next

Cleanup

  • In the notebook: spark.sql("DROP TABLE noaa_bronze_daily")
  • In the notebook: spark.sql("DROP TABLE noaa_silver_daily")
  • Or delete the workspace per Tutorial 01 §6

Troubleshooting

  • Notebook can't read the path: verify the compute identity has Storage Blob Data Contributor on the workspace ADLS account
  • Empty container tree: confirm the Console UAMI has Storage Blob Data Reader on the storage account behind the lakehouse