Configuring a dimension source
- Add a source for the required IBM connection.
- Select the “Dimension” object type.
- Select the dimension and its hierarchy.
- Open the preview and check roots, child elements and attributes.
- Pass the table to a file destination, filter or reference lookup in the flow.
UBC Data Bridge reads attribute definitions, elements and relationships from the selected hierarchy. This section covers reading a dimension. To update a dimension, configure it separately as a destination.
Technical output columns
| Column | Value |
|---|---|
__ibm_dimension | Dimension name |
__ibm_hierarchy | Selected hierarchy name |
__ibm_element | Principal element name |
__ibm_parent | Parent element for this occurrence; empty for a root |
__ibm_depth | Depth from the root, starting at 0 |
__ibm_path | Path as a JSON array of names |
__ibm_weight | Weight of the relationship to the parent |
__ibm_element_type | Numeric, String or Consolidated |
__ibm_consolidation_level | IBM consolidation level |
Hierarchy attributes follow the technical columns: numeric, text and Alias. If an attribute name conflicts with a technical field, its header receives the prefix Attribute:.
Multiple parents and level meanings
An element belonging to several consolidations appears once for each occurrence. The hierarchy row count can therefore exceed the number of unique elements. Use the path and parent element to distinguish occurrences.
Depth from the root and the TM1 consolidation level describe different properties. For example, a leaf element has consolidation level 0 but may be at depth 3. The path is a JSON array, so separator characters in a name do not distort its structure.
Validation and limits
Export validates element types, parent and child existence, weights, duplicate relationships and absence of cycles. An invalid IBM response stops processing. Structure and attributes are subject to the source limit.
To verify, select one consolidation and check its children, order, weights and attributes. If you need a list of unique elements, explicitly remove duplicates after reading using column __ibm_element; doing so intentionally discards information about other parent relationships.