Connect to ClickHouse from Python with clickhouse-connect and run queries, DDL commands, and inserts, including against ClickHouse Cloud.

Documented steps

1. Install the driver: `pip install clickhouse-connect` (add extras as needed, e.g. `pip install "clickhouse-connect[pandas]"` or `[arrow]`, `[async]`, `[sqlalchemy]`).
2. Connect to a local/self-managed server: `import clickhouse_connect; client = clickhouse_connect.get_client(host="localhost", username="default", **{"password": os.environ["CLICKHOUSE_PASSWORD"]})`. The HTTP port is typically 8123 without TLS or 8443 with TLS; pass `port=` explicitly if not using the default.
3. Connect to a ClickHouse Cloud service (requires TLS): `client = clickhouse_connect.get_client(host="HOSTNAME.clickhouse.cloud", port=8443, username="default", **{"password": os.environ["CLICKHOUSE_PASSWORD"]})` — get the hostname/port from the Cloud console's Connect dialog.
4. Run DDL or other non-query statements with `client.command(...)`, e.g. `client.command("CREATE TABLE new_table (key UInt32, value String, metric Float64) ENGINE MergeTree ORDER BY key")`.
5. Insert rows with `client.insert(table_name, data, column_names=[...])`, where data is a 2D array of row values, e.g. `client.insert("new_table", [[1000, "a", 5.233], [2000, "b", -107.04]], column_names=["key", "value", "metric"])`.
6. Run SELECT queries with `client.query(...)` and read results from the `.result_rows` attribute: `result = client.query("SELECT max(key), avg(metric) FROM new_table"); print(result.result_rows)`.
7. Close the client when finished: `client.close()`.
8. Reference: https://clickhouse.com/docs/integrations/python