Database-backed feeds store their physical definition on the Database tab of the Data Vault Source metadata editor. This is not a separate Hop metadata type; it is embedded inside each Data Vault Source alongside the General tab (record source indicator and group).
The database definition tells the model what columns are available, which form the natural or primary key, and what their data types and lengths are.
Why the field layout matters
Accurate column metadata is used throughout the plugin:
-
Load from source (hub Keys tab and satellite Attributes tab) reads primary-key and attribute definitions from the source.
-
Pipeline generation uses primary-key flags to build SELECT lists and hash inputs.
-
Target table DDL uses length and precision so you get realistic column types (VARCHAR(50), DECIMAL(15,2), etc.) instead of generic large types.
Options
| Option | Description |
|---|---|
Database connection |
The Hop database connection (DatabaseMeta) for the source system. Required for Import and for generated pipelines to read data. |
Schema |
Optional schema or catalog that qualifies the table in the source database. |
Source table / view |
Physical table or view name. Can be left empty when the table name is supplied at runtime via variables or when you use the definition only for its field layout. |
Fields (row layout) |
Expected columns from this source. Each field includes: * Field name (as in the source). * Optional description. * Source data type (VARCHAR, NUMBER, DATE, etc.). * Length and precision. * Part of source primary key – Fields marked as primary key become hub business keys and parent-hash inputs for satellites and links. * Hop data type (used internally when building row layouts). The Fields grid is the authoritative column metadata for this feed. |
The Import button
The fastest way to populate the Fields list:
-
Select or enter a database connection.
-
Optionally enter schema and table name.
-
Click Import.
The editor connects to the database, reads column metadata, and attempts to detect primary key columns. It replaces the current Fields list with the imported layout. You can edit the grid afterwards; saved changes become the definition the model uses even if the live source table changes later.
The model editor toolbar Import sources action can create new Data Vault Source objects from database tables in bulk, using the same import machinery.
How the chain works at load time
When a hub, link, or satellite loads data:
-
The table references a Data Vault Source by name (hub: Record sources tab; link/satellite: Default record source).
-
The source’s Database tab supplies connection, table, and field layout.
-
Primary-key fields drive hash calculations; lengths and precisions flow into target DDL.
Tips
-
Import early and verify primary key flags. Wrong PK metadata is the most common cause of incorrect hash keys.
-
Adjust lengths in the field grid when live database metadata is wider than you want in the vault.
-
When the same physical table is loaded with different record source values, create one Data Vault Source per feed (duplicate and change General tab indicator/group settings) rather than sharing one logical name.
-
Use meaningful source names on the General tab; the Database tab can point at the same connection and table for multiple feeds when needed.