DataWorks:MaxCompute table data

The Table Basic Information section on the left side of the details page shows the following metrics:

• View count: Number of views of the table's details page in Data Map over the past 30 days. T+1 delay.

• Read count: Number of times the table is read by production tasks over the past 30 days. Counted task types: SQL, Tunnel Download, data integration, and API-based data services. Only scheduling-associated tasks are counted; reads from tasks not created in DataWorks are excluded. T+1 delay.

• Favorite count: Number of users who have favorited the table. Updated in near real-time.

• Output task: ID of the DataWorks recurring schedule task that writes data to this table. If the table updates periodically but no task ID appears, the updates may originate from tasks outside DataWorks—contact the table owner for details. T+1 delay.

  If you don't have permission to view the output task's code, ask the workspace administrator to grant access. See Enable secure isolation for code and logs.

• Storage size: Logical storage size of the table. T+1 delay.

• Health Status: The rating level derived from the table's governance health score under Data Health > Data Asset Governance.

• Table description: Description of the table. Use Copilot to intelligently generate a description.

Click Details to access field information, partition information, and change records.

Data lineage is generated in real time by parsing actual data flows—scheduled jobs and data synchronization jobs. It shows upstream and downstream dependencies at both the table and field level, letting you trace data origins, identify final destinations, and analyze impact before making changes.

Lineage is available in DataWorks Standard Edition and later.

Lineage from ad hoc queries or manual operations is not captured.

If lineage is missing for SQL run by a PyODPS node, manually configure DataWorks scheduling parameters in the PyODPS task code. See Develop a PyODPS 3 task and Develop a PyODPS 2 task.

Table lineage

The Table Lineage tab shows the full lineage graph for the current table:

• Each node displays upstream and downstream counts. Hover over a table or task type to see its basic information.

• Click a node to search its downstream tables by keyword or by @account_name.

• Click + or - to expand or collapse lineage levels.

Field lineage

The Field Lineage tab shows column-level lineage:

• Switch between fields using the Change Field selector.

• Hover over a field or task type to see its details.

• Click + or - to expand or collapse lineage levels.

Impact analysis

Use Impact Analysis to identify downstream tables affected by a schema or data change before making it. Filter results by lineage level, task type, or table type, then download the analysis.

Impact analysis supports up to 50 lineage levels.

The Data Health section covers two areas: governance health and data quality.

Governance health gives you a quantitative view of the table's asset quality across five dimensions: storage, computing, development, quality, and security. A higher score means better governance.

• Health score: The overall score. Higher is better.

• Pending governance issue trend: A historical view of how pending issues have changed over time.

• Governance items: The list of current issues. Click an issue name or the action button to go to Data Asset Governance > Overview > Governance Work Items and resolve governance items.

Data quality shows the monitoring rules configured for the table and any active DQC alerts. Click Configure Rules to set up or update quality monitoring rules.

Usage records help answer two common questions: who is using this table, and how frequently?

Frequently Associated shows the number of users who are using the current table's data. This is based on the number of times the table is used as a join condition in computations over the past 30 days. T+1 delay.

Access Statistics breaks down usage through charts:

• Read trend chart: Daily read counts for the past 30 days, split by development and production environments. The count reflects how many times each field appears in task code multiplied by how many times the task runs—for example, a field referenced once in a task that runs twice counts as 2; a field referenced twice in a task that runs once also counts as 2. T+1 delay.

• Field references in clauses: How many times each field is used in SQL clauses (WHERE, SELECT, JOIN, GROUP BY). T+1 delay.

• Top readers: The users—tenant accounts used in scheduling and individual accounts—who have read the table via SQL in the past 30 days. T+1 delay.

Data insights analyze the data structure and values to show statistics and distributions. Create a new data insight directly on the table details page, or go to the DataAnalysis module to run data insights.