databricks-hello-world（数据块你好世界）

Databricks Hello World

Overview

Create your first Databricks cluster and notebook via the REST API and Python SDK. Covers single-node dev clusters, SQL warehouses, notebook upload, one-time job runs, and Delta Lake smoke tests.

Prerequisites

• Completed databricks-install-auth setup
• Workspace access with cluster creation permissions
• Valid API credentials in env vars or ~/.databrickscfg

Instructions

Step 1: Create a Single-Node Dev Cluster

# POST /api/2.0/clusters/create
databricks clusters create --json '{
  "cluster_name": "hello-world-dev",
  "spark_version": "14.3.x-scala2.12",
  "node_type_id": "i3.xlarge",
  "autotermination_minutes": 30,
  "num_workers": 0,
  "spark_conf": {
    "spark.databricks.cluster.profile": "singleNode",
    "spark.master": "local[*]"
  },
  "custom_tags": {
    "ResourceClass": "SingleNode"
  }
}'
# Returns: {"cluster_id": "0123-456789-abcde123"}

Or via Python SDK:

from databricks.sdk import WorkspaceClient
from databricks.sdk.service.compute import AutoScale

w = WorkspaceClient()

# create_and_wait blocks until cluster reaches RUNNING state
cluster = w.clusters.create_and_wait(
    cluster_name="hello-world-dev",
    spark_version="14.3.x-scala2.12",
    node_type_id="i3.xlarge",
    num_workers=0,
    autotermination_minutes=30,
    spark_conf={
        "spark.databricks.cluster.profile": "singleNode",
        "spark.master": "local[*]",
    },
)
print(f"Cluster ready: {cluster.cluster_id} ({cluster.state})")

Step 2: Create and Upload a Notebook

import base64
from databricks.sdk import WorkspaceClient
from databricks.sdk.service.workspace import ImportFormat, Language

w = WorkspaceClient()

notebook_source = """
# Databricks notebook source
# COMMAND ----------

# Simple DataFrame
data = [("Alice", 28), ("Bob", 35), ("Charlie", 42)]
df = spark.createDataFrame(data, ["name", "age"])
display(df)

# COMMAND ----------

# Write as Delta table
df.write.format("delta").mode("overwrite").saveAsTable("default.hello_world")

# COMMAND ----------

# Read it back and verify
result = spark.table("default.hello_world")
display(result)
assert result.count() == 3, "Expected 3 rows"
print("Hello from Databricks!")
"""

me = w.current_user.me()
notebook_path = f"/Users/{me.user_name}/hello_world"

w.workspace.import_(
    path=notebook_path,
    format=ImportFormat.SOURCE,
    language=Language.PYTHON,
    content=base64.b64encode(notebook_source.encode()).decode(),
    overwrite=True,
)
print(f"Notebook created at: {notebook_path}")

Step 3: Run the Notebook as a One-Time Job

from databricks.sdk import WorkspaceClient
from databricks.sdk.service.jobs import SubmitTask, NotebookTask

w = WorkspaceClient()

# POST /api/2.1/jobs/runs/submit — no persistent job definition needed
run = w.jobs.submit(
    run_name="hello-world-run",
    tasks=[
        SubmitTask(
            task_key="hello",
            existing_cluster_id="0123-456789-abcde123",  # from Step 1
            notebook_task=NotebookTask(
                notebook_path=f"/Users/{w.current_user.me().user_name}/hello_world",
            ),
        )
    ],
).result()  # .result() blocks until run completes

print(f"Run {run.run_id}: {run.state.result_state}")
# Expect: "Run 12345: SUCCESS"

Step 4: Create a Serverless SQL Warehouse

from databricks.sdk import WorkspaceClient

w = WorkspaceClient()

# Serverless warehouses start in seconds and cost ~$0.07/DBU
warehouse = w.warehouses.create_and_wait(
    name="hello-warehouse",
    cluster_size="2X-Small",
    auto_stop_mins=10,
    warehouse_type="PRO",
    enable_serverless_compute=True,
)
print(f"Warehouse ready: {warehouse.id}")

# Run SQL against it
result = w.statement_execution.execute_statement(
    warehouse_id=warehouse.id,
    statement="SELECT current_timestamp() AS now, current_user() AS who",
)
print(result.result.data_array)

Step 5: Verify Everything via CLI

# List clusters
databricks clusters list --output json | jq '.[] | {id: .cluster_id, name: .cluster_name, state: .state}'

# List workspace contents
databricks workspace list /Users/$(databricks current-user me --output json | jq -r .userName)/

# Get run results
databricks runs list --limit 5 --output json | jq '.runs[] | {run_id: .run_id, name: .run_name, state: .state.result_state}'

# Clean up — terminate the dev cluster (saves money)
databricks clusters delete --cluster-id 0123-456789-abcde123

Output

• Single-node development cluster created and running
• Hello world notebook uploaded to workspace
• Successful notebook execution via runs/submit API
• Serverless SQL warehouse operational
• Delta table default.hello_world created

Error Handling

Examples

Quick Node Type Discovery

w = WorkspaceClient()
# Find cheapest general-purpose instance types
node_types = w.clusters.list_node_types()
for nt in sorted(node_types.node_types, key=lambda x: x.memory_mb)[:5]:
    print(f"{nt.node_type_id}: {nt.memory_mb}MB RAM, {nt.num_cores} cores")

List Available Spark Versions

w = WorkspaceClient()
for v in w.clusters.spark_versions().versions:
    if "LTS" in v.name:
        print(f"{v.key}: {v.name}")

Resources

• Clusters API
• Jobs API — Submit Run
• Workspace API
• SQL Statement Execution API

Next Steps

Proceed to databricks-local-dev-loop for local development setup.