from spicepy import Client

client = Client("YOUR_API_KEY")
reader = client.query("SELECT * FROM my_table LIMIT 10")
df = reader.read_pandas()
print(df)

Choosing HTTP vs. Arrow Flight

HTTP API

Arrow Flight API

Format

JSON

Apache Arrow (binary)

Best for

Simple queries, small results

Large datasets, production workloads

Row limits

Yes

Streaming

Yes

Performance

Good

Best

SDK support

All SDKs

Recommendation: Use Arrow Flight (via SDKs) for production workloads and large result sets. Use the HTTP API for quick testing, small queries, or REST-based integrations.

Query Best Practices

Use `LIMIT` and `OFFSET` for large results

SELECT * FROM my_table
ORDER BY id
LIMIT 1000 OFFSET 0

Increment OFFSET to page through results. Always include ORDER BY for deterministic pagination.

Use recent tables for near-real-time data

Recent tables provide fast access to the last ~30 minutes of data, ideal for dashboards and monitoring.

Combine SQL with client-side processing

Use SQL for filtering, aggregation, and joins, then use client libraries (pandas, NumPy, etc.) for further processing:

reader = client.query("SELECT date, value FROM metrics WHERE date > '2025-01-01'")
df = reader.read_pandas()
df['rolling_avg'] = df['value'].rolling(7).mean()

Use data acceleration for repeated queries

Enable data acceleration on frequently queried datasets to avoid hitting the source on every request.

Common Issues

Arrow Flight TLS errors

Set the GRPC_DEFAULT_SSL_ROOTS_FILE_PATH environment variable:

# macOS
export GRPC_DEFAULT_SSL_ROOTS_FILE_PATH=/etc/ssl/cert.pem

HTTP API returns truncated results

The HTTP API has built-in row and timeout limits. Switch to Arrow Flight or an SDK for unlimited streaming results.

Management API returns `401`

The Management API uses personal access tokens, not app API keys. Generate a token under Profile → Personal Access Tokens.

hashtagAPI Endpoints

hashtagRuntime APIs

hashtagManagement API

hashtagSDKs

hashtagQuick example (Python)

hashtagChoosing HTTP vs. Arrow Flight

hashtagQuery Best Practices

hashtagUse LIMIT and OFFSET for large results

hashtagUse recent tables for near-real-time data

hashtagCombine SQL with client-side processing

hashtagUse data acceleration for repeated queries

hashtagCommon Issues

hashtagArrow Flight TLS errors

hashtagHTTP API returns truncated results

hashtagManagement API returns 401

hashtagFurther Reading

API Endpoints

Runtime APIs

Management API

SDKs

Quick example (Python)