Dynamic Table Scheduling and Deployment Specification

Scheduling Method Selection

For production environments, prioritize using Studio Task to schedule REFRESH:

• Can orchestrate dependencies with upstream and downstream tasks
• Unified monitoring and alerting
• Flexible control over refresh timing and retry strategies

Scheduling Configuration Requirements

Self-Dependency (Required)

• Tasks must configure self-dependency to ensure only one REFRESH instance runs at a time
• Avoid write conflicts or data inconsistency caused by concurrent REFRESH

Upstream Dependencies

• If the source table data needs to be synchronized and ready before refresh, the DT's REFRESH Task should depend on the source table's output task
• If source table data does not require synchronization, upstream dependencies may be omitted

Task Content

Non-Partitioned DT

REFRESH DYNAMIC TABLE schema.dt_name;

Partitioned DT (with Parameters)

SET dt.args.pt = '2024-11-13'; REFRESH DYNAMIC TABLE schema.dt_name PARTITION (pt = '2024-11-13');

The value of dt.args.pt is replaced by the Studio scheduling engine with the specific business date at each execution.

Notes

• REFRESH for different partitions can execute in parallel (assigned to different Tasks or different instances of the same Task)
• Concurrent REFRESH of the same partition or non-partitioned DT is prohibited