column-filtering

Register columnFilteringFeature and createFilteredRowModel through tableFeatures, then add the filter functions your columns reference. Set filterFn on each column definition, such as filterFn: 'includesString' for string columns.

Import individual filterFn_* built-ins like filterFn_includesString and register them in the filterFns option of tableFeatures. Registering only the functions you reference keeps the bundle smaller than importing the full registry.

Built-in string and numeric filters expect comparable accessor values, not objects or rendered UI nodes. Use helper.accessor('status', ...) with a direct key instead of returning objects like { label: row.status } from the accessor.

With manualFiltering: true, the table returns the pre-filtered row model and expects you to send filter state to your server. With manualFiltering: false, the filteredRowModel processes rows client-side; do not combine both approaches.

Controlled filter callbacks receive either a value or an updater function of previous state. Wrap the callback with functionalUpdate(updater, columnFilters) so both forms are handled correctly.

Use filterFromLeafRows with maxLeafRowFilterDepth only when a parent row should survive filtering because a descendant matches. It changes filtering to evaluate leaf rows first in nested or grouped data.