Server-Side Data
Refresh, purge, and inspect a server-owned datasource at runtime.
Use rowModelType="server" when the server owns filtering, sorting, grouping, and range loading. See
Row Models for the full request/response contract; this page covers
the runtime controls once a server datasource is wired up. For pivot mode, the datasource also implements getPivot;
see Server-side pivoting.
<Grid rowModelType='server' columns={columns} datasource={datasource} getRowId={(row: { id: string }) => row.id} />Refreshing and purging
// Swap the datasource entirely (e.g. after auth or endpoint change)
api.setServerSideDatasource(nextDatasource);
// Re-fetch the current view without dropping loaded rows
api.refreshServerSide();
// Re-fetch just one group's route
api.refreshServerSide({ route: ['north-america', 'usa'] });
// Drop cached rows for a route and re-fetch from scratch
api.refreshServerSide({ route: ['north-america'], purge: true });
// Drop cached rows without re-fetching
api.purgeServerSide({ route: ['north-america'] });refreshServerSide and purgeServerSide both accept an optional route — an array of group keys identifying
which store to act on. Omit it to act on the root store.
Inspecting store state
const stores = api.getServerSideStoreState();
// readonly ServerSideStoreSnapshot[] — one entry per route, each with:
// storeId, route, level, rowCountState, blockCount, loadingBlockCount, failedBlockCount, childStoreCountUse this to build a loading/error indicator without re-deriving it from row data — a failed block increments
failedBlockCount on its store snapshot, and api.watch((grid) => grid.status.serverSide, …) hears whenever any of this
changes.
Errors in getRows
Throwing inside your datasource's getRows marks that block failed — the grid does not retry automatically. Surface the failure in your own UI
(via getServerSideStoreState() or the serverSideStateChanged event) and call refreshServerSide({route}) to retry.
Last updated on