Skip to content
+

Data Grid - Server-side tree data

Tree data lazy-loading with server-side data source.

To dynamically load tree data from the server, including lazy-loading of children, you must create a data source and pass the unstable_dataSource prop to the Data Grid, as detailed in the overview section.

The data source also requires some additional props to handle tree data, namely getGroupKey and getChildrenCount. If the children count is not available for some reason, but there are some children, getChildrenCount should return -1.

const customDataSource: GridDataSource = {
  getRows: async (params) => {
    // Fetch the data from the server
  },
  getGroupKey: (row) => {
    // Return the group key for the row, e.g. `name`
    return row.name;
  },
  getChildrenCount: (row) => {
    // Return the number of children for the row
    return row.childrenCount;
  },
};

The following tree data example supports filtering, sorting, and pagination on the server. It also caches the data by default.

Press Enter to start editing

Error handling

For each row group expansion, the data source is called to fetch the children. If an error occurs during the fetch, the grid will display an error message in the row group cell. unstable_onDataSourceError is also triggered with the error and the fetch params.

The demo below shows a toast apart from the default error message in the grouping cell. Cache has been disabled in this demo for simplicity.

Group expansion

The idea behind the group expansion is the same as explained in the Row grouping section. The difference is that the data is not initially available and is fetched automatically after the Data Grid is mounted based on the props defaultGroupingExpansionDepth and isGroupExpandedByDefault in a waterfall manner.

The following demo uses defaultGroupingExpansionDepth='-1' to expand all levels of the tree by default.

Custom cache

The data source uses a cache by default to store the fetched data. Use the unstable_dataSourceCache prop to provide a custom cache if necessary. See Data caching for more info.

The following demo uses QueryClient from @tanstack/react-core as a data source cache.

Press Enter to start editing