Catalog management
The Catalog lets you browse and manage table metadata in your MaxCompute projects. Use its tree view to navigate projects, schemas, tables, and columns. You can also inspect table structures and DDL, preview data samples, restore historical versions with Time Travel, and analyze table health and feature distributions using data insights.
Features
Feature category | Features |
Catalog browsing | Catalog tree, table search, and multi-connection switching |
Table details | Table structure, DDL, and data preview |
Time Travel | Version query, version timeline, and historical version restoration |
Snapshot | Snapshot creation, viewing, and deletion |
Data insights | Volume trend, data profile, quality analysis, anomaly detection, change tracking, attribution analysis, drift detection, correlation analysis, and storage suggestions |
Prerequisites
You have created and connected to a MaxCompute data source. The Catalog loads automatically after a successful connection.
Catalog tree
In the sidebar, click Catalog to open the metadata browsing panel.
Tree structure
The structure of the Catalog tree depends on whether the project uses a two-level or three-level model:
Two-level model: project → table group → table
Three-level model: project → schema → table group → table
Expand a table node to view its column information, including name, type, comment, and whether it is a partition column.
Search
Use the search bar at the top of the panel to search by table or column name. The search returns a flat list of results with a match count.
Right-click actions
Right-click a node in the tree to perform the following actions:
Copy Name: Copy the fully qualified name of the table.
View Details: Open the Table Details panel.
Generate SELECT Query: Generate a SELECT statement for the table in the editor.
Table details
In the Catalog tree, click a table name to open the Table Details panel, which includes the following tabs:
Schema
This tab displays the table's basic properties and column definitions:
Table properties: name, owner, size, row count, creation time, last modified time, lifecycle, table format, whether it is a Delta table, primary key, and storage tier.
Column definitions: column name, data type, comment, and whether it is a partition column.
DDL
This tab displays the table's DDL statement, which you can copy with a single click.
Data preview
This tab loads sample data from the table (100 rows by default). For a partitioned table, you must specify a partition value before previewing. You can refresh to reload the data.
Time Travel
The Time Travel tab in the Table Details panel provides version history and rollback capabilities. This feature is available for Delta Tables and Transactional Tables. For standard tables, you can also view historical versions that are within their backup retention period.
Version query
Use the version offset slider to select a historical version (from the current version back to the earliest version within the retention period). Click Execute Query to run a Time Travel SQL query in the editor and view the data snapshot for that version.
Version timeline
This timeline displays the table's change history. Each entry includes:
Operation type: Operations such as CREATE, INSERT, INSERT INTO, and INSERT OVERWRITE are color-coded for easy identification.
Time: The time when the operation occurred.
LSN: The Log Sequence Number, which uniquely identifies each version.
Each entry supports the following actions:
Preview: View the data for that version.
Copy SQL: Copy the Time Travel query statement.
Restore to this version: Restore the table to the version corresponding to the LSN.
Retention period
The available range for Time Travel depends on the project's backup data retention days, set by odps.timemachine.retention.days (default: 1 day, maximum: 30 days). Delta Tables have a separate acidDataRetainHours setting (default: 168 hours / 7 days). Historical versions beyond the retention period are marked as expired and cannot be restored.
Snapshot
Snapshot management is available under the Time Travel tab in the Table Details panel. A snapshot is a read-only copy of a table at a specific point in time. It is independent of the Time Travel retention period, making it ideal for the long-term preservation of critical versions. This feature supports standard tables and Delta Tables but not Transactional Tables, views, or external tables.
Create snapshot
In the snapshot section, click Create Snapshot.
Enter a snapshot name (default format:
table_name_snap_timestamp).Set the lifecycle (1 to 3,650 days; default: 30 days).
Click Confirm. The system creates the snapshot table.
View and delete snapshots
The snapshot list shows all snapshots for the table. You can delete any snapshot from this list. A deleted snapshot cannot be recovered.
Data insights
The Insight tab in the Table Details panel performs a multi-dimensional analysis of the table to reveal its data characteristics and health.
Dimension | Description |
Volume trend | Shows historical trends of table size and row count. |
Data profile | Provides statistical summaries for each column, such as type distribution, null value rate, and unique value count. |
Data quality | Evaluates quality metrics for nulls, duplicates, and formatting anomalies. |
Anomaly detection | Identifies outliers and abnormal patterns in the data. |
Change tracking | Tracks recent changes in data volume and distribution. |
Attribution analysis | Identifies the likely causes of data changes. |
Drift detection | Analyzes trends in how data distribution shifts over time. |
Correlation analysis | Measures the degree of association between columns. |
Storage suggestions | Provides storage optimization suggestions based on data characteristics. |