Skip to main content
WebGPUnot applicable

TabularGraph

From v9.2

TabularGraph adapts tabular datasets to the Graph interface. It accepts a TabularGraphSource with iterable node and edge handles plus accessor functions for reading and writing state back to the underlying records.

Storage strategy​

Unlike ClassicGraph, TabularGraph stores node and edge state in internal tables that are rebuilt whenever the source version changes. The tables capture:

  • identifiers, directional metadata, and connectivity;
  • logical state (hover, selected, etc.);
  • copies of mutable data blobs exposed through the accessor interface.

Nodes and edges returned from getNodes() / getEdges() expose a index property of type NodeIndex / EdgeIndex. The graph instance provides helper accessors such as getNodeStateByIndex() and getEdgeDataByIndex() for reading or mutating the cached records without holding onto the node or edge objects themselves.

Usage notes​

  • Implement TabularGraphSource so that version increases whenever the underlying dataset changes. The graph rebuilds its tables lazily the next time it is accessed.
  • Provide setState / setData accessors when you need updates made through the runtime to be reflected back into your data source.
  • When working with large immutable datasets, TabularGraph avoids copying handles into new Node and Edge instances while still exposing the Graph API expected by the rendering layers.