Drag tables and files into a notebook
The notebook's Explorer > Data items lists the attached lakehouse's tables and files. As in Fabric, you can put them into your code three ways. Each only writes code into the notebook: nothing is read from OneLake until you run the cell, and then the read uses capacity as any OneLake read does.
Click#
Clicking a table or file adds a cell below the current one that loads it with the lakehouse
module (Polars), as before.
Drag into a cell#
Drag a table or file onto a cell's code:
| Where you drop it | What goes in |
|---|---|
| An empty line of a Python cell | The code that loads it: df = lakehouse.read_table("dbo.sales") and df |
| Inside Python code | The quoted name, "dbo.sales" or "Files/raw/orders.csv" |
| Inside a Python string | The name without quotes |
A %%sql cell |
The table name, dbo.sales |
| A Markdown cell | The name as code, `dbo.sales` |
| Between cells or below the last one | A new cell that loads it |
Ctrl+Z in the cell takes a drop back. A frozen cell doesn't take drops.
Right-click for more#
Right-click a table or file for the ways to load it. Each adds a new cell:
| Table | File (.csv, .parquet, .json) |
|---|---|
Load with Polars (lakehouse.read_table) |
Load with Polars (lakehouse.read_csv / read_parquet / read_json) |
Load with Spark (spark.read.table) |
Load with Spark (spark.read.csv with a header / parquet / json) |
Query with SQL (%%sql SELECT * ... LIMIT 1000) |
Load with pandas (pd.read_csv / read_parquet / read_json on the file's local copy) |
Load with pandas (lakehouse.read_table(...).to_pandas()) |
Read the file's text (not Parquet) |
| Copy name | Copy path |
Spark here is the notebook's local spark (see Fabric notebook code on this PC).
pandas is part of the app's Python engine, as it is in Fabric and Databricks notebooks. A table loads
through Polars and converts with .to_pandas(); a file is read from the local copy that
lakehouse.local_path downloads from Files/.
Need a hand?
Find a fix in troubleshooting, or share a reproducible issue with the community.
Get support ↗Improve this guide on GitHub ↗