MaxCompute:Tenant-level information schema

Tenant-level vs. project-level Information Schema

Use the Tenant-level Information Schema when you need a unified view across projects — for example, auditing access permissions, tracking job resource consumption, or analyzing data usage patterns at the tenant level.

How it works

The Tenant-level Information Schema provides two categories of views:

• Metadata views — browse and retrieve metadata such as tables, columns, schemas, roles, and permissions across all projects.

• Usage history views — analyze job runtime metrics including resource consumption, runtime duration, and data processing volume to optimize jobs or plan compute capacity.

Each view has a defined data freshness level and retention period. Data older than the retention period is no longer available. To retain historical data beyond 14 days or back up job history using snapshots, periodically export data from INFORMATION_SCHEMA to local tables or specified projects.

Limitations

• No cross-Metadata Center access. The Information Schema only covers the Metadata Center where the queried region is located. Metadata Centers are grouped as follows:

  • All regions in the Chinese mainland (Shanghai, Hangzhou, Beijing, Shenzhen, Chengdu, Zhangjiakou, and Ulanqab) form one Metadata Center.

  • Hong Kong (China) forms one Metadata Center.

  • Each region outside China forms one Metadata Center.

  • Finance Cloud (Shanghai Finance Zone and Shenzhen Finance Zone) forms one Metadata Center.

  • Alibaba Gov Cloud (Beijing) forms one Metadata Center.

• `SYSTEM_CATALOG` is read-only. Run queries from projects in the same region as the data you want to query.

• SQL access only. Access views through MaxCompute SQL analysis, the latest version of the client, DataStudio in DataWorks, or SQL scheduling nodes.

• Near-real-time data. Metadata views have latency of approximately 3 hours. For applications that require high metadata timeliness, use the SDK or CLI to retrieve metadata for specific objects directly.

Billing

Compute fees

• Pay-as-you-go projects: Querying Information Schema views incurs standard query fees, calculated based on the underlying expanded SQL. Views are optimized using range clustered tables to reduce input data volume and query costs. For TASKS_HISTORY and TUNNELS_HISTORY, query the previous day's data after 06:00 daily and avoid querying the current day's data — this minimizes input volume and reduces fees.

• Subscription projects: Queries consume purchased compute units (CUs).

Storage fees

Information Schema views do not incur storage fees.

View list

Query examples

Run queries from an SQL-executable project in the same region as your data.

Query table metadata across all projects:

-- Switch to an SQL-executable project in your tenant.
USE <project_name>;
SET odps.namespace.schema=true;
SELECT * FROM SYSTEM_CATALOG.INFORMATION_SCHEMA.tables;

Query job history with a date filter:

-- Switch to an SQL-executable project in your tenant.
USE <project_name>;
SET odps.namespace.schema=true;
SELECT * FROM SYSTEM_CATALOG.INFORMATION_SCHEMA.tasks_history
WHERE ds = 'yyyymmdd'
LIMIT 100;

Grant RAM users access

By default, only the Alibaba Cloud account (main account) can access Information Schema data. To allow a Resource Access Management (RAM) user (sub-account) to query these views, create a tenant-level role with the appropriate policy and assign it to the RAM user.

1. Log on to the MaxCompute console and select a region.MaxCompute console

2. In the left navigation pane, choose Manage Configurations > Tenants.

3. On the Tenants page, click the Roles tab.

4. Click Add Role. In the dialog box, enter a Role Name and Policy Content, then click OK. The following policy grants access to all views under SYSTEM_CATALOG.INFORMATION_SCHEMA:

   {
       "Statement": [
           {
               "Action": [
                   "odps:Describe",
                   "odps:Select"
               ],
               "Effect": "Allow",
               "Resource": [
                   "acs:odps:*:catalogs/system_catalog/schemas/information_schema/tables/*"
               ]
           },
           {
               "Action": [
                   "odps:List"
               ],
               "Effect": "Allow",
               "Resource": [
                   "acs:odps:*:catalogs/system_catalog/schemas/information_schema"
               ]
           }
       ],
       "Version": "1"
   }

5. On the Tenants page, click the Users tab.

6. Click Add Member. In the dialog box, select or add the accounts to assign the role to. In the Batch Set Roles area, select the role you created, click , then click OK. For details on user management, see User Planning and Management.

Usage notes

CATALOGS

List of projects.

CATALOG_PRIVILEGES

Project-level permissions, such as CreateInstance and ListTable.

COLUMNS

Column information for tables in each project.

COLUMN_DATA_MASKING_POLICIES

Data masking information for table columns in each project.

COLUMN_LABEL_GRANTS

Label authorization information for table columns in each project.

COLUMN_LABELS

Column-level label information for tables in each project.

COLUMN_PRIVILEGES

Column-level permission information for tables in each project.

DATA_MASKING_POLICIES

Data masking rule information for each project.

INSTALLED_PACKAGES

Installed packages in each project.

PACKAGE_PRIVILEGES

Package authorization information.

PACKAGE_OBJECTS

Objects in packages.

PARTITIONS

Table partition information for each project.

PARTITION_ACCESS_INFO

Partition access statistics for each project.

RESOURCES

Resource information for each project.

RESOURCE_PRIVILEGES

Resource permission information for each project.

ROLES

Project-level and account-level roles.

SCHEMAS

Schema information for projects.

TABLES

Table information for each project.

TABLE_ACCESS_INFO

Table access statistics for each project.

TABLE_LABEL_GRANTS

Table-level label authorization information.

TABLE_LABELS

Label information for tables in each project.

TABLE_PRIVILEGES

Table permission information for each project.

UDF_PRIVILEGES

UDF authorization information for each project.

UDFS

UDF information for each project.

UDF_RESOURCES

Resource dependencies for UDFs in each project.

USERS

List of users.

USER_ROLES

Role assignments for users.

VOLUMES

MaxCompute volume information. Approximately 3-hour latency.

FOREIGN_SERVERS

Foreign server information. Approximately 3-hour latency.3 hours

QUOTA_USAGE

Real-time snapshots of subscription compute quota usage for monitoring compute resource consumption.

TASKS

Real-time snapshots of currently running jobs. Second-level latency.

TASKS_HISTORY

Completed job history across all projects. Partitioned by date, retaining data for the last 14 days with approximately 3-hour latency.

All ext_* fields require the job submitter to actively pass this information when initiating a job. For field conventions, see Job Common Identifier Conventions.

TUNNELS_HISTORY

Tunnel upload and download history. Partitioned by date, retaining data for the last 14 days.