Skip to content

Version 26.07.1.0

What's new

Impact analysis for RMD elements

Before you change a measure, a dimension or a fact, you can now see everything in the project that the change would affect. The [Open impact analysis] option sits next to [View lineage] for measures, dimensions and facts in the "Measures", "Dimensions" and "Facts" tables of the RMD and in the "Details from RMD" dialog, and it opens an interactive diagram of the selected element together with everything that depends on it, organized by dependency levels — direct dependents first, then the elements that depend on them, and so on. Arrows point from the element towards what it affects, which is the opposite of the lineage view; otherwise the diagram uses the same visual style and interaction patterns as lineage. Dependents include other calculated RMD elements whose formulas reference the element, the fact tables, dimension groups and common measures the element takes part in, the data marts that select it, and the data mart views created in the database for those data marts.

Elements are grouped by level, by type and by the container they belong to. Clicking any RMD element highlights every direct use of it on all levels, together with the connection lines, and a layer selector lets you show or hide complete fact tables and data marts and data mart views in the database; very large trees open partially collapsed. From a node you can open the "Details from RMD" dialog, edit a measure, a fact table or a dimension group, edit a data mart, or open the data mart view management menu — after saving, the tree is rebuilt to reflect the new dependencies. Circular dependencies stop the branch from expanding and are marked with a warning that shows the repeating path on request, and references to elements that no longer exist are also marked. All roles can open impact analysis for the projects they have access to, and the diagram shows only the elements the user is allowed to see.

Change the RMD through the public API

External systems can now update DataForge configurations instead of only reading them, so RMD maintenance can be driven automatically from the other products in your IT environment. Version 2 of the public API adds write operations across every RMD-related area: projects and project versions, measures, dimensions and facts (including bulk creation and update of many elements in a single request), dimension groups together with their dimension membership and hierarchy levels, fact tables together with the measures, dimensions, facts, dimension groups and verification filters assigned to them, and relationships between fact tables and dimension groups. Requests are authenticated with the same API keys as the read operations and exchange JSON.

Incoming data is validated before anything is changed: unknown or invalid fields, names that are already in use, formulas that would create a circular dependency, and deletions of objects still referenced elsewhere are all rejected with an explicit error, and the response codes distinguish blocked access, missing objects, conflicts and business-rule violations. A key can only change data inside its own company, and explicit write access to the target project is required — read access is not sufficient. Every write operation, successful or failed, is recorded in the system journal with the endpoint, the method, the affected object, the caller and the timestamp, and updates additionally record the values before and after the change. Write requests are rate-limited.

Improvements

  • Data mart views can now be managed from the list of data marts, without opening the data mart itself: a new "View in the DB" column shows the name of the view created for each data mart (without the connection, database or schema name) and can be sorted and filtered like any other text column, and the [Manage data mart view in the database] action lets you show, update or delete the view directly from the list (IMP-2833)
  • In on-premises installations, outdated product images are now removed automatically once newer ones are released, so there is no need to clean them up by hand to free disk space (IMP-2834)

Bug fixes

Manage data marts

  • Fixed: hiding elements in a data mart no longer fails (IMP-2829)

Manage fact table details

  • Fixed: database table values are no longer missing in the places where those tables are used (IMP-2738)
  • Fixed: all data is now transferred into a fact table when it is populated or repopulated (IMP-2741)

Manage formulas and SQL scripts

  • Fixed: derived dimensions of a dimension group are now included in the generated script when all dimensions of that dimension group are added to a data mart with the UNION merge type (IMP-2734)
  • Fixed: formula syntax errors are now reported at cell level on all RMD tabs, not only duplicate-name errors (IMP-2821)

Manage lineage and dependencies for elements

  • Fixed: the [View lineage] button is no longer disabled for constant facts (IMP-2835)

Manage project versions

  • Fixed: filters that are not linked to fact tables in the data model are now shown in a newly created version (IMP-2735)

Manage projects

  • Fixed: a project with a problematic connection can now be cloned, and a notification is shown after any project is cloned successfully (IMP-2852)

Manage RMD table content

  • Fixed: copied data is now pasted into the whole selected area of the RMD instead of only the first cell (IMP-2740)
  • Fixed: undoing actions in the RMD now works both from the button above the table and with Ctrl+Z (IMP-2848)

Manage RMDs

  • Fixed: adding a new comment in the RMD no longer opens all the existing comments (IMP-2679)
  • Fixed: the focus on cells and rows is now kept after closing the "Formula builder" and "View lineage" dialogs (IMP-2822)

Manage users and user roles

  • Fixed: the Analyst and Viewer roles can now move nodes in the lineage view and expand formulas in the formula builder (IMP-2847)

Product UI and UX

  • Fixed: an extra scroll bar is no longer shown on the measure verification screen in fact tables and in the data model (IMP-2800)