Dataphin:Manage the metadata inventory

Permissions

Super administrators, system administrators, and custom global roles with the Metadata Discovery - View permission can manage the metadata inventory and view metadata details.

Limitations

Version 1.0 of the PolarDB-X (formerly DRDS) data source does not support collecting view information.

Manage the metadata inventory

1. In the top navigation bar of the Dataphin homepage, choose Governance > Metadata.

2. In the left-side navigation pane, choose Metadata Management > Metadata Discovery to go to the Metadata Discovery page.

3. On the Metadata Discovery page, you can click the Data Tables tab or the Dashboards tab to view metadata for different objects. The Dashboards tab displays metadata for dashboards within a specific application system. You can switch between systems by using the application system selector in the upper-right corner of the page.

   • Data Tables: You can view metadata from the **All Metadata**, **Data Source**, or **Source System** views.

     • When you are in the All Metadata view, the list displays the metadata name, description, collection data source and its code, source system, metadata type, latest version, collection task, and update frequency.

     • When you select a specific Data Source, the list displays the metadata name, Database/Schema/workspace, data source type, data source, description, metadata type, and latest version. You can click View Details next to the list title or View Details next to the data source name to view information about the collection task for that data source, including the data source name, data source code, collection task, update frequency, and source system.

     • When you select a specific Source System, the list displays the metadata name, collection data source and its code, metadata type, latest version, collection task, and update frequency. You can click View Details next to the list title or View Details next to the source system name to view the configuration information for that source system, including the source system name, system type, description, operations owner, and system integration contact.

   • Dashboards: This tab is available only when dashboard metadata is collected. You can view metadata from the **All Dashboards** and **Workspace/Directory** views.

4. (Optional) The directory tree on the left provides different filtering methods depending on the object type.

   • Data table objects: You can switch between different views to search for metadata, or use the search and filter options in the upper-right corner to find specific metadata.

     • Data Source View: Browse metadata based on the source data sources configured for collection. The directory tree displays data source types and data sources. You can expand a data source type and select a directory to automatically filter the list on the right to show metadata belonging to that directory.

     • Source System View: Browse metadata based on the source system that owns the data. You can select a source system to automatically filter the list on the right to show metadata belonging to that source system.

     • Search and Filter: You can perform a fuzzy search by metadata name or description, quickly filter for Deleted Data, or filter by source data source, DB/Schema, source system, or metadata type. You can also filter from different views.

       • From the Data Source View, you can filter by DB/Schema and metadata type.

       • From the Source System View, you can filter by source data source, DB/Schema, and metadata type.

     • All Metadata: You can click All Metadata to view all metadata.

   • Dashboard objects: You can search for metadata from the **All Dashboards** or **Workspace** view, or use the search and filter options in the upper-right corner.

     • Workspace View: Browse metadata based on the source application systems configured for collection. You can select a workspace to automatically filter the list on the right to show dashboards belonging to that workspace.

     • All Dashboards: You can click All Dashboards to view all metadata for the specified application system.

     • Search and Filter: You can perform a fuzzy search by dashboard name or description, or filter by workspace/directory (this filter is available only in the **All Dashboards** view) or source dataset.

5. You can perform the following operations on a target metadata object.

   Actions	Description
   View Metadata	This operation is available only for data table assets. Clicking the name of the collection data source switches to the **Data Source View** and highlights that data source. Clicking the name of the source system switches to the **Source System View** and highlights that source system.
   View Collection Task	This operation is available only for data table assets. You can click the name of the collection task to view its configuration details. For dashboard objects, you can click the details icon next to the application system to go to the corresponding collection task.
   View Metadata Details	You can click the name of the metadata or the View icon in the Operations column to go to the metadata details page. For more information, see View metadata details.
   View Version Details	This operation is available only for data table assets and when the configuration for the data source table is enabled in Metadata > Metadata Change Logs. You can click the Version Management icon in the Operations column to go to the **Version Management** tab of the metadata details page to view information across different versions. For more information, see Version management.
   Delete	If the data is not deleted from the source system, it will reappear after the next collection task runs.

View metadata details

1. Click the name of the target metadata or the **View** icon in the **Operations** column to go to the Metadata Details page.

2. On the Metadata Details page, the information displayed varies depending on the asset type.

   Data tables

   This tab displays the basic information, table details, field details, partition details, and version management for the metadata.

   • Basic Information: Displays the metadata name, data source type, application system type (for DataWorks only), application system name (for DataWorks only), data source and its code, tenant mode (for OceanBase only), Schema/Database/MaxCompute project, metadata type, and collection task information.

     View Collection Task: Click a collection task name to open the collection tasks list, where that task is automatically selected.

   • Table Details: Displays system properties for tables and views (including materialized views), such as business properties, technical properties, and management properties (for DataWorks only).

     • Business Properties: Includes the source system, DataWorks category (for DataWorks only), DataWorks tags (for DataWorks only), and DataWorks usage instructions (for DataWorks only).

     • Technical Properties: Includes the description, whether it is a lakehouse table, lakehouse format, storage type, storage size, and other information. The information that can be collected varies by data source. For more information, see Metadata properties.

       When the data source is DataWorks, it displays the data source type, data source name, project name, database name/schema name, description, creation time, last DDL update time, last data change time, whether it is a partitioned table, storage size, and lifecycle.

       Note

       • For Oracle tables, the storage size is an estimate based on the row count and the average space used per row (in bytes). It does not include statistics for indexes, unused space, or fragmentation. The statistical value will be lower than the actual space occupied.

       • If the data source table is a partitioned table, the row count is usually an estimate and is for reference only. Whether it is updated depends on the metadata update mechanism of each data source system. For example, in MySQL or OceanBase, you may need to manually run the analyze command to update the count.

       • Currently, collecting the row count for the main table of a partitioned table is not supported.

       • For an OceanBase data source (Oracle tenant), you must run the analyze command to collect the table row count. The command is as follows:

         -- Replace the table name in the curly braces ({}) with the name of your data table
         analyze table {schema.tablename}  COMPUTE STATISTICS;

       • For data sources that support creating lakehouse tables, such as Hive, the system determines whether it is a lakehouse table and its format based on the collected information.

     • Management Properties (for DataWorks only): Includes the DataWorks workspace and the DataWorks table owner.

     • DDL Statement: When the metadata type is a view or materialized view, you can view and copy the DDL statement used to create the view or materialized view in the data source.

   • Field Details: If the current table is a partitioned table, you can view partition information. The information displayed varies for some data sources. For more information, see Metadata properties.

     You can search by field name or description, and switch between different versions to view field details.

   • Partition Details: Partition information can be collected for only some data source types. For more information, see Metadata properties.

     Multi-level partitions are displayed with levels separated by forward slashes (/).

     Note

     The record count and storage size for partitions are for reference only.

   • Version Management: You can view the version history of the metadata. You can modify the number of viewable versions in Metadata > Metadata Change Logs.

     The system generates a new version when it detects changes in the source system metadata during two consecutive collections, based on the collection policy of the collection task. For information about how versions are generated for different data sources, see Metadata versions.

     For example, if the data update policy is set to Add new metadata and update changed metadata and Ignore delete operations, a new metadata version will not be generated if, during two consecutive collections, the source system only has deleted data but no new or updated data.

     • You can search for versions by name.

     • The system displays the version number, changes from the previous version, table description, and version creation time. You can also view a version comparison.

       • Changes from previous version: Compares the source system metadata changes against the previous version, based on the collection policy configured in the collection task.

       • Version Comparison: Compare historical versions. The system displays basic information, changed fields, and all fields. For more information, see the Version comparison section below.

   Version comparison

   On the Version Comparison page, you can switch between versions to compare metadata changes.

   • Version Selection: Select a base version and a comparison version to view the differences in metadata between them.

     • Base Version: You can only select a historical version that was generated before the comparison version.

     • Comparison Version: You can only select a new version that was generated after the base version.

   • Basic Information: Displays the collection completion time, execution method, DDL statement (viewable for views/materialized views), and the table properties that affect version generation. For information about table properties, see Metadata versions.

     For DataWorks assets, it displays the metadata type, field count, table description, whether it is a partitioned table, DataWorks category, DataWorks tags, DataWorks usage instructions, and DataWorks table owner.

   • Changed Fields: Displays the differences between the selected comparison version and the base version, including the sequence number (not displayed for DataWorks assets), field name, data type, nullability (not displayed for DataWorks assets), default value (not displayed for DataWorks assets), description, and change type.

     For example, if the field's description was modified in the comparison version, the change type for that field is 'update'.

   • All Fields: Displays all field information for both the base version and the comparison version, including the sequence number (not displayed for DataWorks assets), field name, data type, nullability (not displayed for DataWorks assets), default value (not displayed for DataWorks assets), and description.

   Dashboards

   This tab displays the basic information, properties, and chart information for the dashboard.

   • Basic Information: Displays the dashboard name, application system type, application system, and collection task information.

     • View Collection Task: Click a collection task name to open the collection tasks list, where that task is automatically selected.

     • View Metadata: You can quickly search for and switch to view the details of other metadata.

     • View Dashboard: Click View Dashboard to open the dashboard's details page in Quick BI.

   • Properties: Displays the properties of the dashboard, including business properties, management properties, and technical properties.

     • Business Properties: Displays the description of the dashboard.

     • Management Properties: Displays the BI workspace, directory, owner, last updater, last update time, collaboration policy, public access status, and public access expiration time.

       Public Access: Public reports can be accessed directly via a link without logging in, while non-public reports require you to log in to Quick BI.

     • Technical Properties: Displays the source dataset name, total number of charts, and third-party embedding status.

   • Chart Information: Displays the most recently collected chart information, including the chart sequence number, chart name, chart type, field type, and field name/data type.

     You can search by chart name or filter by chart type, dimension, or measure.