Examples¶

Runnable examples live in the examples/ directory of the repository, organized by language. Each one connects to a Spark Connect server at sc://localhost:15002 by default; override the URI to point at your own endpoint.

All examples assume a running Apache Spark Connect server. See the Quickstart for how to start one. The Python examples are executed against live Spark 3.5.x, 4.0.x, and 4.1.x servers on every CI run.

Python¶

In examples/python/:

File	What it shows
`01_quickstart.py`	Connect, run a `SELECT`, and print rows.
`02_arrow_and_pandas.py`	`fetch_arrow_table()` (zero-copy) and `fetch_df()` into pandas.
`03_parameters.py`	Prepared statements with positional `?` (`qmark`) parameters.
`04_metadata.py`	List catalogs, schemas, tables, columns, and fetch a table schema.
`05_ddl_and_dml.py`	`CREATE` / `INSERT` via `cursor.execute`, then `SELECT`.
`06_auth_tls.py`	Connect with a bearer `token=` over TLS (Databricks-style).
`07_polars.py`	Read straight into a Polars DataFrame with `pl.read_database`.
`08_duckdb.py`	Query Spark's Arrow output in DuckDB, zero-copy.
`09_pyarrow_streaming.py`	Stream a large result as a pyarrow `RecordBatchReader`.
`10_pandas_read_sql.py`	`pandas.read_sql` over the ADBC connection.

Examples 07 to 10 are covered in depth on the Ecosystem Integrations page.

Run one with:

pip install adbc-driver-spark pyarrow pandas
python examples/python/01_quickstart.py

Go¶

In examples/go/:

Directory	What it shows
`quickstart/`	`NewDriver`, open a database, run a query, iterate the `RecordReader`.
`metadata/`	`GetObjects`, `GetTableSchema`, `GetTableTypes`.
`parameters/`	`Prepare` plus `Bind` for parameter binding.

Run one with:

go run ./examples/go/quickstart

C¶

In examples/c/:

File	What it shows
`quickstart.c`	Connect, run a `SELECT`, and read the Arrow result stream (counts rows and batches).
`README.md`	Build, run, and going-further notes.

quickstart.c loads libadbc_driver_spark through the standard ADBC driver manager (libadbc_driver_manager): set the driver database option to the shared library path and the manager dlopen()s it and resolves the default AdbcDriverInit entrypoint. See Using from C and C++ for the full listing and the compile and run commands.

R¶

In examples/r/:

File	What it shows
`quickstart.R`	Connect, run a `SELECT`, and materialize the Arrow result as a `data.frame`.
`README.md`	Setup and run instructions.

quickstart.R loads the shared library through the standard ADBC driver manager for R, adbcdrivermanager: adbc_driver() wraps libadbc_driver_spark and the manager resolves the default AdbcDriverInit entrypoint. See Using from R for the full listing and run commands.

Rust¶

In examples/rust/: a Cargo project that loads libadbc_driver_spark through the adbc_driver_manager crate and reads the Arrow result. See Using from Rust.

Ruby¶

In examples/ruby/: a script that loads libadbc_driver_spark through Red ADBC (the adbc gem) and prints the Arrow table. See Using from Ruby.

Tip

Set the URI through an environment variable so the same example works against a local server and a remote one, for example SPARK_CONNECT_URI=sc://spark.example.com:443.