Data Model
  • 29 Nov 2022
  • 2 Minutes to read
  • Dark
    Light

Data Model

  • Dark
    Light

Article summary

The Data Model zone provides users a dynamic way to interact with the objects, measures, and reports managed by the platform. These data assets can be displayed as a two-dimensional, "dependency tree" graph or as a list of assets. In the "dependency tree" view, users can browse connections between an object and its immediate upstream or downstream relatives by clicking on the nodes of the graph. Both views can be toggled for any lens via an icon-button at the top of the screen, but each lens will default to its most suitable view when initially selected. When users are already in a report, measure, or object, navigation to the data model screen from the leftnav will automatically put the focus on the current asset.

The search bar helps users to filter by object name, table name, and object type. They can use the same bar to perform a “deep search,” which will comb through and return results for matched strings in measure or object definitions. Deep search is a one-stop tool to find all references for a given string in the metadata for any object in the library.

Lenses

Lenses are used to highlight the breadth of layer-spanning functions and as a way to visualize overall data flow throughout the system. For example, the Recent ELT Timing lens allows users to select one of the most recent ELT jobs and shows the relative build times of each object. The default lens upon first entry to the screen shows a user their own recent activity.

The “Field Name Use” lens can be used to determine what the effects of renaming an upstream field may be. If the field is merely referenced in a restriction or derived field of a downstream object, then that object will be included in the lens, but the inclusion will not trigger any propagation to further downstream objects. On the other hand, if the field is published the object will both be included and will also propagate the trace to its downstream objects. Bespoke SQL objects, by their nature, are difficult to parse exactly. For the purposes of the lens, a bespoke SQL object will both be included and will propagate to its children if it publishes a field with the exact name of the field being searched. Reports will be included in the lens if they contain boards which reference the field, either as a split or a filter. News feed posts are not considered triggers for inclusion in the lens, even if they reference the searched field.

The “Field Use” lens is similar to the “Field Name Use” lens, but is more expansive in its scope, because it will track downstream usage even in renamed fields. For example, if an upstream field is an input to a derived field in a downstream object, that downstream field will itself be traced on its further use.


Was this article helpful?