Datus MaxCompute Adapter

Alibaba Cloud MaxCompute adapter for Datus. It supports both legacy project.table projects and schema-enabled project.schema.table projects.

Configuration

database:
  type: maxcompute
  database: ${MAXCOMPUTE_PROJECT}
  endpoint: ${MAXCOMPUTE_ENDPOINT}
  access_key_id: ${MAXCOMPUTE_ACCESS_KEY_ID}
  access_key_secret: ${MAXCOMPUTE_ACCESS_KEY_SECRET}
  namespace_mode: auto

For a three-level project, schema is optional and defaults to default. Successful automatic detection is cached per connector. Errors other than the specific MaxCompute response for a non-three-level project are propagated.

namespace_mode accepts auto, two_level, or three_level. Use an explicit mode only when automatic probing is unavailable to the configured identity. Optional settings include quota_name, tunnel_endpoint, timeout_seconds, query_timeout_seconds, and default_hints.

Testing

Unit tests use mocked PyODPS clients and do not require a MaxCompute service:

uv run --package datus-maxcompute pytest datus-maxcompute/tests/unit

MaxCompute is a managed cloud service, so this adapter does not rely on a local Docker service. Real two-level and three-level coverage lives in the MaxCompute Cloud Tests workflow. It is not scheduled or required by normal CI; run it manually after configuring the protected maxcompute-integration environment with these secrets:

• MAXCOMPUTE_ENDPOINT
• MAXCOMPUTE_ACCESS_KEY_ID
• MAXCOMPUTE_ACCESS_KEY_SECRET
• MAXCOMPUTE_TWO_LEVEL_PROJECT
• MAXCOMPUTE_THREE_LEVEL_PROJECT