Tutorial: Synapse notebook editor¶
CSA Loom
synapse-notebookeditor — the Synapse Studio-style Spark notebook designer: multi-language cells on a Synapse Big Data pool via Livy, authored over the Synapse dev plane. No Microsoft Fabric required.
What it is¶
A Synapse notebook is the Spark authoring surface in Synapse Studio — multi-language cells (PySpark, Spark Scala, Spark SQL, SparkR, .NET Spark C#) run interactively on a Synapse Big Data pool via Livy. In Loom it reads/writes the workspace notebook artifact over the Synapse dev plane and runs cells against a live Livy session through the Console managed identity.
When to use it¶
- Interactive Spark data engineering / exploration against the lake with the full Synapse cell model (per-cell language, parameters cell, outline).
- You want notebooks stored as real Synapse workspace artifacts (and backed up to ADLS), runnable later from pipelines.
Step-by-step in Loom¶
- Open the editor. Choose + New item → Notebook and pick the Synapse notebook flavor (or open an existing one). The editor opens at
/items/synapse-notebook/<id>. - Attach a Spark pool. Pick a Big Data pool from the attach picker; the first run cold-starts the session (about 2–3 minutes).
- Attach an environment (optional). Pick a Spark configuration to apply library packages and Spark session settings to the pool — surfaced from the workspace's
sparkconfigurations. - Author cells. Add code or markdown cells between any two cells, set the notebook default language and per-cell language, reorder, duplicate, and collapse cells in the designer.
- Mark a parameters cell. Designate one code cell as the parameters cell so its variables can be overridden when the notebook runs from a pipeline (papermill / ADF).
- Navigate with the outline. The left-panel outline tracks headings from markdown cells; click an entry to scroll to that cell.
- Run and inspect. Run a cell or Run all; output and error tracebacks render inline from the Livy statement result.
- Publish. Save publishes the notebook back to the Synapse workspace as an artifact and backs up the
.ipynbto ADLSsilver/loom/notebooks/.
The Azure backend it rides on¶
- Artifacts: the Synapse dev plane (workspace notebook artifacts).
- Execution: Livy sessions on a Synapse Big Data pool via the Console managed identity.
- Backup:
.ipynbcopies to ADLS Gen2.
No Fabric required¶
Authoring and execution are Synapse + ADLS; no Fabric capacity, workspace, or OneLake is involved.