Metadata management is a module in the administrative system’s Data Fusion framework. It is a place where the metadata tracked by the system is centrally managed, such as modifying the display name of events, modifying the display name of event attributes, modifying the unit of event attributes, and modifying the display name of user attributes.

To ensure flexible use by users and to meet the needs of different data collection requirements, the system provides a Free Data Reporting Mode and a Data Strict Validation Mode.

With the Data Strict Validation Mode, data accuracy can be effectively improved, addressing the issue of erroneous data reporting from the very source. Once this mode is enabled, after updating the events and attributes to be collected into the project, all subsequent data reported and stored will need to be validated against the configured events and attribute types. Only data that meets the requirements will be stored; data that does not will be rejected.

Using the Free Data Reporting Mode allows enterprises to quickly perform data collection tasks, but the data quality entirely depends on the accuracy of the developers' tracking code. In this mode, any event and attribute can be reported at any time. Subsequent data related to continuously reported attributes will be compared and validated against the attribute type and attribute name stored on first entry. If the information does not match the existing database, it will be regarded as a brand new meta-event, event attribute, or user attribute and will be stored in the current project.

In the user table/event table, attribute names must be unique, and data types must be consistent. For example, if there are two meta-events in the event table: registration and login, both having the attribute username, it must be ensured that the type of this username is consistent.