DataWind supports direct integration with Volcano Engine DataFinder (hereinafter referred to as Finder). Behavior analytics data collected by Finder can be analyzed directly, or associated with business data for joint analysis, enabling deeper analytical insights.
Prerequisites:

• For private deployment of DataWind, the corresponding Finder must also be privately deployed. Both products need to be deployed in advance.
• For SaaS DataWind, the corresponding Finder must be SaaS. You need to obtain a Finder account in advance, and the primary account used to apply for DataWind and Finder must be the same primary account.
• Version requirements:
  • Private deployment: Finder must be version V4.3.0 or above, and DataWind must be version V2.50.0 or above.
  • SaaS: Finder must be the SaaS cloud-native version.

2.1 Data connection steps #

(1) Click to enter a specific project, click Data Preparation, find Data Connection in the dropdown list, and click Create Data Connection.

(2) Click Application Data Source on the left, and select "BytePlus Growth Analytics DataFinder".

(3) In the basic information for the data connection, select the corresponding application from the dropdown and click Save.

(4) Click the One-Click Create Dataset button. The system will automatically create the dataset.
(5) In the popup after successful creation, click "View here" to jump to the corresponding dataset and view details.

2.2 Application scope selectable during the data connection process #

• For private deployment, there is a one-to-one relationship between projects. Only one application can be selected under each project space.
• For the SaaS version, all applications created under the primary Volcano account can be selected.

2.3 User permission restrictions #

Who can create a Finder data connection?

• No distinction between private deployment or SaaS
• On Finder, authorization controls the SQL custom query feature. Users authorized to use this feature can create Finder data connections on DataWind; users without authorization cannot.

2.4 Description of integrated table structures #

• DataWind can integrate the events table, users table, and items table from Finder.
• The table structure information for integrated Finder data can be viewed in the dataset details.
• For the mapping of fields to Finder sources, see Data Format.
• For detailed parsing instructions, see the events table, users table, and items table in Field Description.
• When using the event_params.loc_city_id field in the events table, note that this field must be converted before it can be used.
  • Click the dataset's field configuration to go to the field configuration details page, then click Add Field.
  • Enter the following expression in the field expression box and click Save.

dictGetOrDefault('default_user_dictionary', 'target_value', tuple('loc_city_id', ifNull(event_params.loc_city_id, 'null')), 'unknown')

• In subsequent applications of the events table, the city field can be used directly when required.
• Parsed fields in the events table
  The event_params.xxx.yyy field, user_profiles.xxx field, and item_profiles.xxx.yyyy field will be parsed into multiple fields based on the actual reporting and definitions. The following is an example of parsing the event_params.xxx.yyy field.

2.5 Feature limitations #

• The dataset extraction method is direct connection only.
• The dataset does not support join or union operations.
• Visual modeling capability is not supported.