Skip to main content
Version: Current 18.0.0 🎯

Build Your First Topic

Goal​

Define your first Topic in the Admin UI: choose kind/type, attach a data source, add factors (columns), set indexes, and save.

Prerequisites​

  • Admin access to Topic management.
  • At least one data source configured (RDS or NoSQL).

1) Open Topic​

  • Navigate to Admin → Topic.
  • Click Create Topic to open the standard tuple page.

2) Basic Information​

  • Name: snake/camel case; becomes table/document name in the data source.
  • Kind: System / Business / Synonym.
    • Use Business for application data; System for logs/infra data.
    • Synonym maps an existing external table; structure is fixed, read‑only, RDS only.
  • Type: Raw / Meta / Distinct / Aggregation (Aggregate/Time/Ratio).
    • Start with Distinct for instance data (e.g., orders, customers).
    • Choose Raw when receiving JSON via API and storing immutable events.
    • Choose Meta for definitions/reference data used by pipelines.
  • Data Source: pick where the topic will be stored physically.
  • Description: optional.

3) Add Factors​

  • Add one row per attribute with:
    • Name and Label.
    • Type: pick from built‑in types (Number/Text/Boolean/Date/Time and many domain types).
    • Default Value: optional.
    • Encryption: optional. Use AES/MD5/SHA256 or masking for sensitive fields.
  • For Raw topics:
    • Use Object/Array for hierarchical JSON.
    • Use dot notation for nested fields (e.g., customer.name).
    • Tick Flatten Column on selected nested fields to create searchable columns; index flattened fields.

4) Indexing​

  • Assign Index Group for factors you query by; use Unique Index Group when values must be unique.
  • In RDS, indexes avoid full table scans and improve pipeline reads on criteria.

5) Enumeration​

  • When a factor is Enum, link it to an Enumeration so UI and DQC know allowed values.

6) Save​

  • Validation checklist:
    • Topic: Name, Kind, Type, Data Source are required.
    • Factors: at least one with Name and Type.
    • For Raw, Flatten Column is required on any dot‑name factor you mark as flattened.
  • Click Save.

7) Optional: Generate Scripts​

  • Click Download Scripts to export creation/alteration scripts for Oracle/MySQL/MSSQL or Liquibase.
  • Review and version the scripts before executing in your environment.

8) Optional: Build Flatten Topics (from Raw)​

  • Click Build Flatten Topics, choose target distinct topics.
  • The system generates topics and pipelines to copy data from the Raw topic into Distinct topics.

9) Optional: Import Metadata (Meta)​

  • For Meta topics, download CSV/JSON templates and import data.
  • Synonym meta cannot import data.

Good Practices​

  • Index flattened factors and any criteria used in pipelines.
  • Keep names unique within the data source.
  • Use Business kind for application data; Synonym for external tables (read‑only).
  • Encrypt or mask sensitive fields (contact info, IDs, etc.).
  • Design topic layers: Raw → Distinct → Aggregation/Meta, and move data via pipelines.

What’s Next​

  • Create pipelines for the new topic.
  • Build subjects/datasets and reports on top of the topic.