Migrating to TanStack Table V9 (React)

What's New in TanStack Table V9#

TanStack Table V9 is a major release with significant internal architectural improvements while maintaining the core table logic you're familiar with. Here are the key changes:

1. Better Performance#

• Lower memory usage: The core architecture now shares more behavior across table objects, with some large-table scenarios seeing up to 90% memory savings.

• Faster client-side row models: Sorting, filtering, and aggregation paths have improved algorithms and memoization, with many scenarios seeing up to 40-70% speed improvements.

• Better column resizing performance: Column resizing also gets significant performance improvements from the same architectural and memoization work.

2. State Management Overhaul (Backward Compatible)#

• React Compiler compatibility: The state system is built on TanStack Store, giving the table a reactive foundation that works correctly under the React Compiler.

• Fine-grained subscriptions: State slices can be read independently through table.atoms, table.store, table.state, selectors, or table.Subscribe.

• External state or atoms: You can still use state plus on[State]Change, or own individual slices with writable atoms via the new atoms option.

3. Type-Safety Improvements#

• New and revamped type helpers: New type helpers help define columns, custom filters, sorts, aggregations, column and table meta, shared table options and components, and more!

• Per-table meta types: tableMeta, columnMeta, and filterMeta slots let you type meta for a specific table instead of globally augmenting shared interfaces. No more global declaration merging required!

• Feature-gated APIs: APIs only exist when their feature is registered, and tableFeatures() validates feature prerequisites at the type level.

4. Tree Shaking and Extensibility#

• Import only the features you use: Tables that only need sorting do not ship filtering, pagination, or other unused feature code. Start with 5kb of bundled JS and only bundle the features you need.

• Tree-shakeable row models and functions: Row model factories and filterFns / sortFns / aggregationFns now live on tableFeatures(), so unused processing code can be dropped.

• Custom features use the same system: Your own feature plugins can register state, options, and APIs alongside the built-in features. See the Custom Features Guide.

5. Composability#

• tableOptions: Compose reusable table configuration, including features, row models, and default options.

• createTableHook: Create custom table hooks with pre-bound features and components when you need a reusable app-level table pattern. See the composable-tables (createTableHook) guide.

The Good News: Most Upgrades Are Opt-in#

While Table V9 is a significant upgrade, you don't have to adopt everything at once:

• Don't want to optimize renders yet? Do nothing special. The default selector selects all registered state, so rendering works like Table V8.

• Don't want to think about tree-shaking? Import stockFeatures to include all features, just like Table V8.

• Table markup is largely unchanged. How you render <table>, <thead>, <tr>, <td>, etc. remains the same.

  The main change is how you define a table with the useTable hook, specifically the new features option and where row model factories are registered.

Quick Legacy Migration#

Need to migrate incrementally? We are providing a temporary shortcut with the useLegacyTable hook. It accepts the Table V8-style API while using Table V9 under the hood. This is deprecated and intended only as a temporary migration aid. It includes all features by default, resulting in a larger bundle size than you even got with Table V8.

Legacy APIs live in a separate export. Import core utilities from @tanstack/react-table and legacy-specific APIs from @tanstack/react-table/legacy:

import { flexRender } from '@tanstack/react-table'
import {
  useLegacyTable,
  getCoreRowModel,
  getFilteredRowModel,
  getSortedRowModel,
  getPaginationRowModel,
  legacyCreateColumnHelper,
} from '@tanstack/react-table/legacy'

See the useLegacyTable Guide for full documentation, examples, and type helpers.

The rest of this guide focuses on migrating to the full Table V9 API and taking advantage of its features.

Core Breaking Changes#

Hook Rename#

The hook name has been simplified to be consistent across all TanStack libraries:

// Table V8
import { useReactTable } from '@tanstack/react-table'
const table = useReactTable(options)

// Table V9
import { useTable } from '@tanstack/react-table'
const table = useTable(options)

Instance Methods Must Be Called on Their Instance#

In Table V9, methods on rows, cells, columns, headers, and similar table objects are shared on the object's prototype instead of being created as arrow functions on each object. This improves memory usage, but it means destructuring those methods loses the this context they need to operate on the instance.

// Table V8 - worked because getValue closed over the row object
const { getValue } = row
const value = getValue('name')

// Table V9 - call the method on the instance
const value = row.getValue('name')

This applies to row, cell, column, header, and related instance APIs, but not to the table instance itself. Audit code that destructures methods from table objects or passes them around as bare callbacks. Prefer calling them through the original object, for example row.getValue('name'), cell.getContext(), column.getCanSort(), or header.getContext().

Because these methods now live on the prototype, they also do not appear as own properties in Object.keys(instance), object spread, or JSON.stringify. A shallow clone like { ...row } copies row data but does not copy row methods. The methods are still callable normally because JavaScript looks them up through the prototype chain.

New Required features Table Option#

In Table V9, you must explicitly declare which features your table uses. Features, Row Models, and Row Model processing "Fns" are defined on the new features table option.

In Table V8, all features were bundled and included in the useReactTable hook. In Table V9, you import only what you need.

// Table V8
import {
  useReactTable,
  getCoreRowModel,
  getSortedRowModel,
  sortingFns,
} from '@tanstack/react-table'

const table = useReactTable({
  columns,
  data,
  getCoreRowModel: getCoreRowModel(),
  getSortedRowModel: getSortedRowModel(),
  sortingFns,
})

// Table V9
import {
  useTable,
  tableFeatures,
  rowSortingFeature,
  createSortedRowModel,
  sortFns,
} from '@tanstack/react-table'

// All table options that concern including code modules (features, row models, Fns, etc.)
const features = tableFeatures({
  rowSortingFeature, // new - import and pass the feature you want to use
  sortedRowModel: createSortedRowModel(), // now row models are defined on the features object
  sortFns, // now Fns are defined on the features object
  // ...more features, row models, etc.
})

const table = useTable({
  features, // new required option
  columns,
  data,
})

Shortcut: Use stockFeatures for Table V8-like Behavior#

If you want all features without having to think about it (like Table V8), import stockFeatures:

import { useTable, stockFeatures } from '@tanstack/react-table'

const table = useTable({
  features: stockFeatures, // All features included - just like Table V8 (though larger bundle now)
  columns,
  data,
})

Available Features#

Row Model Factories#

Row models are the functions that process your data (filtering, sorting, pagination, etc.). In Table V9, row model factories live on the tableFeatures({}) call rather than a separate rowModels option. The processing function registries (filterFns, sortFns, aggregationFns) are also registered on features. Row model slots are type-checked, so each row model must be specified after its associated feature in the same tableFeatures call.

Migration Mapping#

Key Change: Row Model Factories and Fns Registries Move to tableFeatures#

Row model factories and their processing function registries are now slots on tableFeatures. This enables better tree-shaking: you only bundle the row model code and filter/sort/aggregation functions you actually register.

import {
  tableFeatures,
  createFilteredRowModel,
  createSortedRowModel,
  createGroupedRowModel,
  filterFns, // Built-in filter functions
  sortFns, // Built-in sort functions
  aggregationFns, // Built-in aggregation functions
} from '@tanstack/react-table'

const features = tableFeatures({
  columnFilteringFeature,
  rowSortingFeature,
  columnGroupingFeature,
  rowPaginationFeature,
  filteredRowModel: createFilteredRowModel(),
  sortedRowModel: createSortedRowModel(),
  groupedRowModel: createGroupedRowModel(),
  paginatedRowModel: createPaginatedRowModel(),
  filterFns,
  sortFns,
  aggregationFns,
})

const table = useTable({
  features,
  columns,
  data,
})

State Management Changes#

Table V9's state system is built on TanStack Store and exposes three read surfaces on the table instance:

Writable counterparts (mostly internal):

Accessing State#

In Table V8, you accessed state via table.getState(). In Table V9, state is accessed differently:

// Table V8
const state = table.getState()
const { sorting, pagination } = table.getState()

// Table V9 (RECOMMENDED) - via table.state (full selected state by default)
const table = useTable({
  features,
  columns,
  data,
})
const { sorting, pagination } = table.state

// Table V9 - via the store (full state)
const fullState = table.store.state
const { sorting, pagination } = table.store.state

// Table V9 - via table.state with a custom selector
const selectedTable = useTable(options, (state) => ({
  sorting: state.sorting,
  pagination: state.pagination,
}))
// Now selectedTable.state only contains sorting and pagination
const { sorting, pagination } = selectedTable.state

// Table V9 - via a single slice atom (framework-agnostic, ideal for fine-grained subscriptions)
const sorting = table.atoms.sorting.get()

onStateChange Table Option Removed#

The onStateChange table option was removed in Table V9, although individual on[State]Change handlers are still available for specific state slices.

If you want to lift or listen to any state change, you can set up a subscription to the table.store

const unsubscribe = table.store.subscribe((state) => {
  console.log(state)
})

Optimized Rendering with table.Subscribe#

The biggest state management improvement is table.Subscribe, which enables fine-grained reactivity:

function MyTable() {
  const table = useTable(
    {
      features,
      columns,
      data,
    },
    (state) => ({}), // default is state => state - this opts out of all default re-renders in favor of manual subscriptions down below
  )

  return (
    <table.Subscribe
      selector={(state) => ({
        sorting: state.sorting,
        pagination: state.pagination,
      })}
    >
      {({ sorting, pagination }) => (
        // This only re-renders when sorting or pagination changes
        <div>
          <table>{/* ... */}</table>
          <div>Page {pagination.pageIndex + 1}</div>
        </div>
      )}
    </table.Subscribe>
  )
}

Default: Table V8-Style Full State Subscription#

The default selector already gives Table V8-style behavior where the component re-renders on any registered table state change:

const table = useTable({
  features,
  columns,
  data,
})

// table.state contains the full registered state
const { sorting, pagination, columnFilters } = table.state

Passing (state) => state is equivalent to the default and is no longer necessary. Pass a custom selector when you want table.state to contain only specific slices, or pass () => null and use table.Subscribe lower in the tree when the parent should not re-render for table state changes.

Controlled State#

The Table V8-style state + on[State]Change controlled state patterns still work and remain convenient for simple integrations. For new Table V9 code, prefer owning state slices with external atoms (see External Atoms below), which give you fine-grained subscriptions without mirroring state through React:

const [sorting, setSorting] = useState<SortingState>([])
const [pagination, setPagination] = useState<PaginationState>({
  pageIndex: 0,
  pageSize: 10,
})

const table = useTable({
  features,
  columns,
  data,
  state: {
    sorting,
    pagination,
  },
  onSortingChange: setSorting,
  onPaginationChange: setPagination,
})

Per-Slice Atom Subscriptions#

Because each state slice is backed by its own atom, you can subscribe a component to a single slice without re-rendering on any other state change. Use useSelector from @tanstack/react-store with table.atoms.<slice>:

import { useSelector } from '@tanstack/react-store'

function PaginationFooter({ table }) {
  // Re-renders only when pagination changes. Sorting, filtering, selection, etc. are all ignored.
  const pagination = useSelector(table.atoms.pagination)

  return <div>Page {pagination.pageIndex + 1}</div>
}

This is the narrowest subscription surface available. Compared to table.Subscribe, which selects from the full table.store.state, reading a per-slice atom skips even constructing the full state snapshot on change.

When to reach for table.atoms vs. table.Subscribe: Both give you fine-grained re-renders. table.Subscribe is nicer when you want to project multiple slices into a single rendered block. table.atoms.<slice> is nicer when a component only cares about one slice, or when you're passing a subscription source to non-table code.

External Atoms (Advanced)#

For advanced patterns (sharing a slice across tables, integrating with atom-based libraries, or wiring a slice up to persistence), Table V9 lets you own individual state slices yourself by passing writable atoms via the new atoms option. See the Basic External Atoms example.

import { useCreateAtom, useSelector } from '@tanstack/react-store'
import {
  useTable,
  tableFeatures,
  rowSortingFeature,
  rowPaginationFeature,
  createSortedRowModel,
  createPaginatedRowModel,
  sortFns,
} from '@tanstack/react-table'
import type { PaginationState, SortingState } from '@tanstack/react-table'

const features = tableFeatures({ rowSortingFeature, rowPaginationFeature })

function MyTable({ data, columns }) {
  // Create stable external atoms for the slices you want to own.
  const sortingAtom = useCreateAtom<SortingState>([])
  const paginationAtom = useCreateAtom<PaginationState>({
    pageIndex: 0,
    pageSize: 10,
  })

  // Subscribe to each atom independently for fine-grained reactivity.
  const sorting = useSelector(sortingAtom)
  const pagination = useSelector(paginationAtom)

  const table = useTable({
    features,
    columns,
    data,
    // Per-slice external atoms. The library writes directly to these,
    // bypassing the internal baseAtoms for those slices.
    atoms: {
      sorting: sortingAtom,
      pagination: paginationAtom,
    },
  })

  // Table writes like table.setPageIndex(2) go straight to `paginationAtom`.
  // Any other subscriber of `paginationAtom` will see the update too.
  // ...
}

How External Atoms Interact with state and on*Change#

When you register an external atom for a slice:

• Reads: The derived table.atoms[slice] and table.store.state[slice] both read from your external atom.

• Writes: Library writes (e.g. table.setSorting(...), column.toggleSorting()) go directly to your external atom's set(). You do not need a corresponding onSortingChange handler; owning the atom is the subscription.

• Precedence: If you pass both options.atoms[key] and options.state[key], the atom wins. If you pass neither, Table V9 falls back to its internal baseAtoms[key] (Table V8-style self-managed state).

• Reset: table.reset() does not clear external atoms. You own them, so you decide when to reset. Call myAtom.set(defaultValue) yourself if needed.

When to Choose External Atoms vs. Controlled State#

Column Helper Changes#

The createColumnHelper function now requires a TFeatures type parameter in addition to TData:

// Table V8
import { createColumnHelper } from '@tanstack/react-table'

const columnHelper = createColumnHelper<Person>()

// Table V9
import {
  createColumnHelper,
  tableFeatures,
  rowSortingFeature,
} from '@tanstack/react-table'

const features = tableFeatures({ rowSortingFeature })
const columnHelper = createColumnHelper<typeof features, Person>()

New columns() Helper Method#

Table V9 adds a columns() helper for better type inference when wrapping column arrays. In Table V8, TValue wasn't always type-safe, especially with group columns, where nested column types could be lost or widened. The columns() helper uses variadic tuple types to preserve each column's individual TValue type, so info.getValue() and cell renderers stay correctly typed throughout nested structures:

const columnHelper = createColumnHelper<typeof features, Person>()

// Wrap your columns array for better type inference
const columns = columnHelper.columns([
  columnHelper.accessor('firstName', {
    header: 'First Name',
    cell: (info) => info.getValue(),
  }),
  columnHelper.accessor('lastName', {
    id: 'lastName',
    header: () => <span>Last Name</span>,
    cell: (info) => <i>{info.getValue()}</i>,
  }),
  columnHelper.display({
    id: 'actions',
    header: 'Actions',
    cell: (info) => <button>Edit</button>,
  }),
])

Rendering Changes#

flexRender Function#

The flexRender function still exists and works the same way:

import { flexRender } from '@tanstack/react-table'

// Still works in Table V9
flexRender(cell.column.columnDef.cell, cell.getContext())
flexRender(header.column.columnDef.header, header.getContext())

New FlexRender and <table.FlexRender /> Component#

Table V9 adds a cleaner component-based approach attached to the table instance:

const table = useTable({ /* ... */ })

// Instead of:
{flexRender(header.column.columnDef.header, header.getContext())}

// You can use:
<table.FlexRender header={header} />
<table.FlexRender cell={cell} />
<table.FlexRender footer={footer} />
// or
import { FlexRender } from '@tanstack/react-table'
<FlexRender header={header} />
<FlexRender cell={cell} />
<FlexRender footer={footer} />

This should be way more convenient than the old flexRender function!

The tableOptions() Utility#

The tableOptions() helper provides type-safe composition of table options. It's useful for creating reusable partial configurations that can be spread into your table setup.

Basic Usage#

import {
  tableOptions,
  tableFeatures,
  rowSortingFeature,
} from '@tanstack/react-table'

const features = tableFeatures({ rowSortingFeature })

// Create a reusable options object with features pre-configured
const baseOptions = tableOptions({
  features,
  debugTable: process.env.NODE_ENV === 'development',
})

// Use in your table; columns, data, and other options can be added
const table = useTable({
  ...baseOptions,
  columns,
  data,
})

Composing Partial Options#

tableOptions() allows you to omit certain required fields (like data, columns, or features) when creating partial configurations:

// Row model factories and fns registries are registered on the features object
const features = tableFeatures({
  rowSortingFeature,
  columnFilteringFeature,
  sortedRowModel: createSortedRowModel(),
  filteredRowModel: createFilteredRowModel(),
  sortFns,
  filterFns,
})

// Partial options without data or columns
const featureOptions = tableOptions({
  features,
  getRowId: (row) => row.id,
  manualPagination: true,
  manualSorting: true,
  manualFiltering: true,
})

// Another partial without features (inherits from spread)
const paginationDefaults = tableOptions({
  initialState: {
    pagination: { pageIndex: 0, pageSize: 25 },
  },
})

// Combine them
const table = useTable({
  ...featureOptions,
  ...paginationDefaults,
  columns,
  data,
})

Using with createTableHook#

tableOptions() pairs well with createTableHook for building composable table factories:

const features = tableFeatures({
  rowSortingFeature,
  rowPaginationFeature,
  sortedRowModel: createSortedRowModel(),
  paginatedRowModel: createPaginatedRowModel(),
  sortFns,
})

const sharedOptions = tableOptions({ features })

const { useAppTable } = createTableHook(sharedOptions)

createTableHook: Composable Table Patterns#

This is an advanced, optional feature. You don't need to use createTableHook; useTable is sufficient for most use cases. If you're familiar with TanStack Form's createFormHook, createTableHook works almost the same way: it creates a custom hook with pre-bound configuration that you can reuse across many tables.

For applications with multiple tables sharing the same configuration, createTableHook lets you define features (including row model factories), and reusable components once:

// hooks/table.ts
import {
  createTableHook,
  tableFeatures,
  columnFilteringFeature,
  rowSortingFeature,
  rowPaginationFeature,
  createFilteredRowModel,
  createSortedRowModel,
  createPaginatedRowModel,
  filterFns,
  sortFns,
} from '@tanstack/react-table'

// Import your reusable components
import { PaginationControls, SortIndicator, TextCell } from './components'

// Features and row model factories defined once
const features = tableFeatures({
  columnFilteringFeature,
  rowSortingFeature,
  rowPaginationFeature,
  filteredRowModel: createFilteredRowModel(),
  sortedRowModel: createSortedRowModel(),
  paginatedRowModel: createPaginatedRowModel(),
  filterFns,
  sortFns,
})

export const {
  useAppTable,
  createAppColumnHelper,
  useTableContext,
  useCellContext,
  useHeaderContext,
} = createTableHook({
  features,

  // Default table options
  debugTable: process.env.NODE_ENV === 'development',

  // Register reusable components
  tableComponents: { PaginationControls },
  cellComponents: { TextCell },
  headerComponents: { SortIndicator },
})

Using useAppTable#

// features/users.tsx
import { useAppTable, createAppColumnHelper } from './hooks/table'

const columnHelper = createAppColumnHelper<Person>()

const columns = columnHelper.columns([
  columnHelper.accessor('firstName', {
    header: 'First Name',
    cell: ({ cell }) => <cell.TextCell />, // Pre-bound component!
  }),
])

function UsersTable({ data }: { data: Person[] }) {
  const table = useAppTable({
    columns,
    data,
    // features (including row model factories) already configured!
  })

  return (
    <table.AppTable>
      <table>
        <thead>
          {table.getHeaderGroups().map((headerGroup) => (
            <tr key={headerGroup.id}>
              {headerGroup.headers.map((h) => (
                <table.AppHeader header={h} key={h.id}>
                  {(header) => (
                    <th>
                      <header.FlexRender />
                      <header.SortIndicator />
                    </th>
                  )}
                </table.AppHeader>
              ))}
            </tr>
          ))}
        </thead>
        <tbody>
          {table.getRowModel().rows.map((row) => (
            <tr key={row.id}>
              {row.getAllCells().map((c) => (
                <table.AppCell cell={c} key={c.id}>
                  {(cell) => (
                    <td>
                      <cell.FlexRender />
                    </td>
                  )}
                </table.AppCell>
              ))}
            </tr>
          ))}
        </tbody>
      </table>
      <table.PaginationControls />
    </table.AppTable>
  )
}

Context Hooks for Components#

Components registered via createTableHook can access their context:

// components/SortIndicator.tsx
import { useHeaderContext } from './hooks/table'

export function SortIndicator() {
  const header = useHeaderContext()
  const sorted = header.column.getIsSorted()

  if (!sorted) return null
  return sorted === 'asc' ? ' 🔼' : ' 🔽'
}

// components/TextCell.tsx
import { useCellContext } from './hooks/table'

export function TextCell() {
  const cell = useCellContext()
  return <span>{cell.getValue() as string}</span>
}

// components/PaginationControls.tsx
import { useTableContext } from './hooks/table'

export function PaginationControls() {
  const table = useTableContext()

  return (
    <table.Subscribe selector={(s) => s.pagination}>
      {(pagination) => (
        <div>
          <button onClick={() => table.previousPage()}>Previous</button>
          <span>Page {pagination.pageIndex + 1}</span>
          <button onClick={() => table.nextPage()}>Next</button>
        </div>
      )}
    </table.Subscribe>
  )
}

Other Breaking Changes#

Column Pinning Option Split#

The enablePinning option has been split into separate options:

// Table V8
enablePinning: true

// Table V9
enableColumnPinning: true
enableRowPinning: true

Removed Internal APIs#

All internal APIs prefixed with _ have been removed. If you were using any of these, use their public equivalents:

• Removed: table._getPinnedRows()

• Removed: table._getFacetedRowModel()

• Removed: table._getFacetedMinMaxValues()

• Removed: table._getFacetedUniqueValues()

Column Sizing vs. Column Resizing Split#

In Table V8, column sizing and resizing were combined in a single feature. In Table V9, they've been split into separate features for better tree-shaking.

If you only need column sizing (fixed widths) without interactive resizing, you can import just columnSizingFeature. If you need drag-to-resize functionality, import both:

import {
  columnSizingFeature,
  columnResizingFeature,
} from '@tanstack/react-table'

const features = tableFeatures({
  columnSizingFeature,
  columnResizingFeature, // Only if you need interactive resizing
})

Sorting API Renames#

Sorting-related APIs have been renamed for consistency:

Update your column definitions:

// Table V8
const columns = [
  {
    accessorKey: 'name',
    sortingFn: 'alphanumeric', // or custom function
  },
]

// Table V9
const columns = [
  {
    accessorKey: 'name',
    sortFn: 'alphanumeric', // or custom function
  },
]

Row API Changes#

Some row APIs have changed from private to public:

If you were accessing this internal API, you can now use it without the underscore prefix.

Row Selection API Changes#

The "some rows selected" checks were simplified to mean "at least one row is selected":

In Table V8 these returned false once every row was selected; in Table V9 they stay true. If you use them to drive an indeterminate "select all" checkbox, gate the indeterminate state on the matching all-selected check so it clears at full selection:

getIsSomeRowsSelected() && !getIsAllRowsSelected()

TypeScript Changes Summary#

Type Generics#

Most types now require a TFeatures parameter:

// Table V8
type Column<TData>
type ColumnDef<TData>
type Table<TData>
type Row<TData>
type Cell<TData, TValue>

// Table V9
type Column<TFeatures, TData, TValue>
type ColumnDef<TFeatures, TData, TValue>
type Table<TFeatures, TData>
type Row<TFeatures, TData>
type Cell<TFeatures, TData, TValue>

Using typeof features#

The easiest way to get the TFeatures type is with typeof:

const features = tableFeatures({
  rowSortingFeature,
  columnFilteringFeature,
})

// Use typeof to get the type
type MyFeatures = typeof features

const columns: ColumnDef<typeof features, Person>[] = [...]

function Filter({ column }: { column: Column<typeof features, Person, unknown> }) {
  // ...
}

Using StockFeatures#

If using stockFeatures with useTable, use the StockFeatures type:

import type { StockFeatures, ColumnDef } from '@tanstack/react-table'

const columns: ColumnDef<StockFeatures, Person>[] = [...]

TableMeta/ColumnMeta Typing Changes#

No more declaration merging required! (Although it still works if you want to keep using it)

Global declaration merging to extend TableMeta or ColumnMeta works exactly like it did in Table V8. The only change you need to make is updating the generics shape: both interfaces now take TFeatures as the first type parameter.

// Table V8
declare module '@tanstack/react-table' {
  interface ColumnMeta<TData, TValue> {
    customProperty: string
  }
}

// Table V9 - TFeatures is now the first parameter
declare module '@tanstack/react-table' {
  interface ColumnMeta<TFeatures, TData, TValue> {
    customProperty: string
  }
}

That's all that's required if you want to keep declaring meta types globally.

Optionally, Table V9 also adds a new way to declare meta types per-table without declaration merging. You can use type-only tableMeta/columnMeta slots on the features option, which only affect tables created with that features object:

const features = tableFeatures({
  rowSortingFeature,
  columnMeta: metaHelper<{ customProperty: string }>(),
})

See the new Table and Column Meta Guide for full details on both approaches.

FilterFns/SortFns/AggregationFns/FilterMeta Augmentation Replaced by Registry Slots#

In Table V8, making a custom function usable as a string reference (like filterFn: 'fuzzy') required declare module augmentation of the FilterFns interface, and typing filter meta required augmenting FilterMeta. In Table V9, registering the function in the matching registry slot does both jobs with no global augmentation:

// Table V8
declare module '@tanstack/react-table' {
  interface FilterFns {
    fuzzy: FilterFn<unknown>
  }
  interface FilterMeta {
    itemRank: RankingInfo
  }
}

// Table V9 - register in the slot; the key becomes a valid string value
interface FuzzyFilterMeta {
  itemRank?: RankingInfo
}

const features = tableFeatures({
  columnFilteringFeature,
  filteredRowModel: createFilteredRowModel(),
  filterFns: { ...filterFns, fuzzy: fuzzyFilter },
  filterMeta: metaHelper<FuzzyFilterMeta>(),
})

// 'fuzzy' now typechecks in column defs for tables using these features
columnHelper.accessor('name', { filterFn: 'fuzzy' })

The same pattern applies to sortFns (for sortFn string values) and aggregationFns (for aggregationFn string values). See the Fuzzy Filtering Guide for a complete example.

RowData Type Restriction#

The RowData type is now more restrictive:

// Table V8 - very permissive
type RowData = unknown

// Table V9 - must be a record or array
type RowData = Record<string, any> | Array<any>

This change improves type safety. If you were passing unusual data types, ensure your data conforms to Record<string, any> or Array<any>.

Migration Checklist#

• Update import: useReactTable → useTable
• Define features using tableFeatures() (or use stockFeatures)
• Migrate get*RowModel() options to tableFeatures slots (e.g. filteredRowModel: createFilteredRowModel())
• Register filterFns / sortFns / aggregationFns registries as slots on tableFeatures (row model factories no longer take arguments)
• Replace declare module augmentation of FilterFns/SortFns/AggregationFns with registry-slot registration, and FilterMeta augmentation with the filterMeta slot
• Update TypeScript types to include TFeatures generic
• Update state access: table.getState() → table.store.state or table.state
• Update createColumnHelper<TData>() → createColumnHelper<TFeatures, TData>()
• Replace enablePinning with enableColumnPinning/enableRowPinning if used
• Rename sortingFn → sortFn in column definitions
• Split column sizing/resizing: use both columnSizingFeature and columnResizingFeature if needed
• Rename columnSizingInfo state → columnResizing (and related options)
• If you use TableMeta/ColumnMeta declaration merging, add the TFeatures generic to your augmentations (optionally, switch to the per-table tableMeta/columnMeta feature slots)
• Replace destructured row/cell/column/header methods with calls on the instance (for example, row.getValue('name'))
• (Optional) Add table.Subscribe for render optimizations
• (Optional) Subscribe to individual slices via table.atoms.<slice> + useSelector for the narrowest re-renders
• (Optional) Pass writable atoms via the new atoms option to own specific state slices externally
• (Optional) Use tableOptions() for composable configurations
• (Optional) Migrate to createTableHook for reusable table patterns

Examples#

Check out these examples to see Table V9 patterns in action:

• Basic useTable - Simple table with the new useTable hook
• Basic useLegacyTable - Migration example using useLegacyTable
• Basic useAppTable - Using createTableHook
• Basic External State - Classic state + on*Change controlled state
• Basic External Atoms - Owning state slices with useCreateAtom + the atoms option
• Filters - Filtering with the new API
• Sorting - Sorting with the new API
• Composable Tables - Advanced createTableHook patterns
• Kitchen Sink - A comprehensive example of all features