home / skills / jezweb / claude-skills / tanstack-query

tanstack-query skill

/skills/tanstack-query

This skill helps you migrate to TanStack Query v5 and debug SSR, hydration, and infinite queries with object-based hooks and advanced options.

npx playbooks add skill jezweb/claude-skills --skill tanstack-query

Review the files below or copy the command above to add this skill to your agents.

Files (23)
SKILL.md
30.5 KB
---
name: tanstack-query
description: |
  Manage server state in React with TanStack Query v5. Covers useMutationState, simplified optimistic updates, throwOnError, network mode (offline/PWA), and infiniteQueryOptions.

  Use when setting up data fetching, fixing v4→v5 migration errors (object syntax, gcTime, isPending, keepPreviousData), or debugging SSR/hydration issues with streaming server components.
user-invocable: true
---

# TanStack Query (React Query) v5

**Last Updated**: 2026-01-20
**Versions**: @tanstack/[email protected], @tanstack/[email protected]
**Requires**: React 18.0+ (useSyncExternalStore), TypeScript 4.7+ (recommended)

---

## v5 New Features

### useMutationState - Cross-Component Mutation Tracking

Access mutation state from anywhere without prop drilling:

```tsx
import { useMutationState } from '@tanstack/react-query'

function GlobalLoadingIndicator() {
  // Get all pending mutations
  const pendingMutations = useMutationState({
    filters: { status: 'pending' },
    select: (mutation) => mutation.state.variables,
  })

  if (pendingMutations.length === 0) return null
  return <div>Saving {pendingMutations.length} items...</div>
}

// Filter by mutation key
const todoMutations = useMutationState({
  filters: { mutationKey: ['addTodo'] },
})
```

### Simplified Optimistic Updates

New pattern using `variables` - no cache manipulation, no rollback needed:

```tsx
function TodoList() {
  const { data: todos } = useQuery({ queryKey: ['todos'], queryFn: fetchTodos })

  const addTodo = useMutation({
    mutationKey: ['addTodo'],
    mutationFn: (newTodo) => api.addTodo(newTodo),
    onSuccess: () => {
      queryClient.invalidateQueries({ queryKey: ['todos'] })
    },
  })

  // Show optimistic UI using variables from pending mutations
  const pendingTodos = useMutationState({
    filters: { mutationKey: ['addTodo'], status: 'pending' },
    select: (mutation) => mutation.state.variables,
  })

  return (
    <ul>
      {todos?.map(todo => <li key={todo.id}>{todo.title}</li>)}
      {/* Show pending items with visual indicator */}
      {pendingTodos.map((todo, i) => (
        <li key={`pending-${i}`} style={{ opacity: 0.5 }}>{todo.title}</li>
      ))}
    </ul>
  )
}
```

### throwOnError - Error Boundaries

Renamed from `useErrorBoundary` (breaking change):

```tsx
import { QueryErrorResetBoundary } from '@tanstack/react-query'
import { ErrorBoundary } from 'react-error-boundary'

function App() {
  return (
    <QueryErrorResetBoundary>
      {({ reset }) => (
        <ErrorBoundary onReset={reset} fallbackRender={({ resetErrorBoundary }) => (
          <div>
            Error! <button onClick={resetErrorBoundary}>Retry</button>
          </div>
        )}>
          <Todos />
        </ErrorBoundary>
      )}
    </QueryErrorResetBoundary>
  )
}

function Todos() {
  const { data } = useQuery({
    queryKey: ['todos'],
    queryFn: fetchTodos,
    throwOnError: true, // ✅ v5 (was useErrorBoundary in v4)
  })
  return <div>{data.map(...)}</div>
}
```

### Network Mode (Offline/PWA Support)

Control behavior when offline:

```tsx
const queryClient = new QueryClient({
  defaultOptions: {
    queries: {
      networkMode: 'offlineFirst', // Use cache when offline
    },
  },
})

// Per-query override
useQuery({
  queryKey: ['todos'],
  queryFn: fetchTodos,
  networkMode: 'always', // Always try, even offline (for local APIs)
})
```

| Mode | Behavior |
|------|----------|
| `online` (default) | Only fetch when online |
| `always` | Always try (useful for local/service worker APIs) |
| `offlineFirst` | Use cache first, fetch when online |

**Detecting paused state:**
```tsx
const { isPending, fetchStatus } = useQuery(...)
// isPending + fetchStatus === 'paused' = offline, waiting for network
```

### useQueries with Combine

Combine results from parallel queries:

```tsx
const results = useQueries({
  queries: userIds.map(id => ({
    queryKey: ['user', id],
    queryFn: () => fetchUser(id),
  })),
  combine: (results) => ({
    data: results.map(r => r.data),
    pending: results.some(r => r.isPending),
    error: results.find(r => r.error)?.error,
  }),
})

// Access combined result
if (results.pending) return <Loading />
console.log(results.data) // [user1, user2, user3]
```

### infiniteQueryOptions Helper

Type-safe factory for infinite queries (parallel to `queryOptions`):

```tsx
import { infiniteQueryOptions, useInfiniteQuery, prefetchInfiniteQuery } from '@tanstack/react-query'

const todosInfiniteOptions = infiniteQueryOptions({
  queryKey: ['todos', 'infinite'],
  queryFn: ({ pageParam }) => fetchTodosPage(pageParam),
  initialPageParam: 0,
  getNextPageParam: (lastPage) => lastPage.nextCursor,
})

// Reuse across hooks
useInfiniteQuery(todosInfiniteOptions)
useSuspenseInfiniteQuery(todosInfiniteOptions)
prefetchInfiniteQuery(queryClient, todosInfiniteOptions)
```

### maxPages - Memory Optimization

Limit pages stored in cache for infinite queries:

```tsx
useInfiniteQuery({
  queryKey: ['posts'],
  queryFn: ({ pageParam }) => fetchPosts(pageParam),
  initialPageParam: 0,
  getNextPageParam: (lastPage) => lastPage.nextCursor,
  getPreviousPageParam: (firstPage) => firstPage.prevCursor, // Required with maxPages
  maxPages: 3, // Only keep 3 pages in memory
})
```

**Note:** `maxPages` requires bi-directional pagination (`getNextPageParam` AND `getPreviousPageParam`).

---

## Quick Setup

```bash
npm install @tanstack/react-query@latest
npm install -D @tanstack/react-query-devtools@latest
```

### Step 2: Provider + Config

```tsx
// src/main.tsx
import { QueryClient, QueryClientProvider } from '@tanstack/react-query'
import { ReactQueryDevtools } from '@tanstack/react-query-devtools'

const queryClient = new QueryClient({
  defaultOptions: {
    queries: {
      staleTime: 1000 * 60 * 5, // 5 min
      gcTime: 1000 * 60 * 60, // 1 hour (v5: renamed from cacheTime)
      refetchOnWindowFocus: false,
    },
  },
})

<QueryClientProvider client={queryClient}>
  <App />
  <ReactQueryDevtools initialIsOpen={false} />
</QueryClientProvider>
```

### Step 3: Query + Mutation Hooks

```tsx
// src/hooks/useTodos.ts
import { useQuery, useMutation, useQueryClient, queryOptions } from '@tanstack/react-query'

// Query options factory (v5 pattern)
export const todosQueryOptions = queryOptions({
  queryKey: ['todos'],
  queryFn: async () => {
    const res = await fetch('/api/todos')
    if (!res.ok) throw new Error('Failed to fetch')
    return res.json()
  },
})

export function useTodos() {
  return useQuery(todosQueryOptions)
}

export function useAddTodo() {
  const queryClient = useQueryClient()
  return useMutation({
    mutationFn: async (newTodo) => {
      const res = await fetch('/api/todos', {
        method: 'POST',
        headers: { 'Content-Type': 'application/json' },
        body: JSON.stringify(newTodo),
      })
      if (!res.ok) throw new Error('Failed to add')
      return res.json()
    },
    onSuccess: () => {
      queryClient.invalidateQueries({ queryKey: ['todos'] })
    },
  })
}

// Usage:
function TodoList() {
  const { data, isPending, isError, error } = useTodos()
  const { mutate } = useAddTodo()

  if (isPending) return <div>Loading...</div>
  if (isError) return <div>Error: {error.message}</div>
  return <ul>{data.map(todo => <li key={todo.id}>{todo.title}</li>)}</ul>
}
```

---

## Critical Rules

### Always Do

✅ **Use object syntax for all hooks**
```tsx
// v5 ONLY supports this:
useQuery({ queryKey, queryFn, ...options })
useMutation({ mutationFn, ...options })
```

✅ **Use array query keys**
```tsx
queryKey: ['todos']              // List
queryKey: ['todos', id]          // Detail
queryKey: ['todos', { filter }]  // Filtered
```

✅ **Configure staleTime appropriately**
```tsx
staleTime: 1000 * 60 * 5 // 5 min - prevents excessive refetches
```

✅ **Use isPending for initial loading state**
```tsx
if (isPending) return <Loading />
// isPending = no data yet AND fetching
```

✅ **Throw errors in queryFn**
```tsx
if (!response.ok) throw new Error('Failed')
```

✅ **Invalidate queries after mutations**
```tsx
onSuccess: () => {
  queryClient.invalidateQueries({ queryKey: ['todos'] })
}
```

✅ **Use queryOptions factory for reusable patterns**
```tsx
const opts = queryOptions({ queryKey, queryFn })
useQuery(opts)
useSuspenseQuery(opts)
prefetchQuery(opts)
```

✅ **Use gcTime (not cacheTime)**
```tsx
gcTime: 1000 * 60 * 60 // 1 hour
```

### Never Do

❌ **Never use v4 array/function syntax**
```tsx
// v4 (removed in v5):
useQuery(['todos'], fetchTodos, options) // ❌

// v5 (correct):
useQuery({ queryKey: ['todos'], queryFn: fetchTodos }) // ✅
```

❌ **Never use query callbacks (onSuccess, onError, onSettled in queries)**
```tsx
// v5 removed these from queries:
useQuery({
  queryKey: ['todos'],
  queryFn: fetchTodos,
  onSuccess: (data) => {}, // ❌ Removed in v5
})

// Use useEffect instead:
const { data } = useQuery({ queryKey: ['todos'], queryFn: fetchTodos })
useEffect(() => {
  if (data) {
    // Do something
  }
}, [data])

// Or use mutation callbacks (still supported):
useMutation({
  mutationFn: addTodo,
  onSuccess: () => {}, // ✅ Still works for mutations
})
```

❌ **Never use deprecated options**
```tsx
// Deprecated in v5:
cacheTime: 1000 // ❌ Use gcTime instead
isLoading: true // ❌ Meaning changed, use isPending
keepPreviousData: true // ❌ Use placeholderData instead
onSuccess: () => {} // ❌ Removed from queries
useErrorBoundary: true // ❌ Use throwOnError instead
```

❌ **Never assume isLoading means "no data yet"**
```tsx
// v5 changed this:
isLoading = isPending && isFetching // ❌ Now means "pending AND fetching"
isPending = no data yet // ✅ Use this for initial load
```

❌ **Never forget initialPageParam for infinite queries**
```tsx
// v5 requires this:
useInfiniteQuery({
  queryKey: ['projects'],
  queryFn: ({ pageParam }) => fetchProjects(pageParam),
  initialPageParam: 0, // ✅ Required in v5
  getNextPageParam: (lastPage) => lastPage.nextCursor,
})
```

❌ **Never use enabled with useSuspenseQuery**
```tsx
// Not allowed:
useSuspenseQuery({
  queryKey: ['todo', id],
  queryFn: () => fetchTodo(id),
  enabled: !!id, // ❌ Not available with suspense
})

// Use conditional rendering instead:
{id && <TodoComponent id={id} />}
```

❌ **Never rely on refetchOnMount: false for errored queries**
```tsx
// Doesn't work - errors are always stale
useQuery({
  queryKey: ['data'],
  queryFn: failingFetch,
  refetchOnMount: false,  // ❌ Ignored when query has error
})

// Use retryOnMount instead
useQuery({
  queryKey: ['data'],
  queryFn: failingFetch,
  refetchOnMount: false,
  retryOnMount: false,  // ✅ Prevents refetch for errored queries
  retry: 0,
})
```

---

## Known Issues Prevention

This skill prevents **16 documented issues** from v5 migration, SSR/hydration bugs, and common mistakes:

### Issue #1: Object Syntax Required
**Error**: `useQuery is not a function` or type errors
**Source**: [v5 Migration Guide](https://tanstack.com/query/latest/docs/framework/react/guides/migrating-to-v5#removed-overloads-in-favor-of-object-syntax)
**Why It Happens**: v5 removed all function overloads, only object syntax works
**Prevention**: Always use `useQuery({ queryKey, queryFn, ...options })`

**Before (v4):**
```tsx
useQuery(['todos'], fetchTodos, { staleTime: 5000 })
```

**After (v5):**
```tsx
useQuery({
  queryKey: ['todos'],
  queryFn: fetchTodos,
  staleTime: 5000
})
```

### Issue #2: Query Callbacks Removed
**Error**: Callbacks don't run, TypeScript errors
**Source**: [v5 Breaking Changes](https://tanstack.com/query/latest/docs/framework/react/guides/migrating-to-v5#callbacks-on-usequery-and-queryobserver-have-been-removed)
**Why It Happens**: onSuccess, onError, onSettled removed from queries (still work in mutations)
**Prevention**: Use `useEffect` for side effects, or move logic to mutation callbacks

**Before (v4):**
```tsx
useQuery({
  queryKey: ['todos'],
  queryFn: fetchTodos,
  onSuccess: (data) => {
    console.log('Todos loaded:', data)
  },
})
```

**After (v5):**
```tsx
const { data } = useQuery({ queryKey: ['todos'], queryFn: fetchTodos })
useEffect(() => {
  if (data) {
    console.log('Todos loaded:', data)
  }
}, [data])
```

### Issue #3: Status Loading → Pending
**Error**: UI shows wrong loading state
**Source**: [v5 Migration: isLoading renamed](https://tanstack.com/query/latest/docs/framework/react/guides/migrating-to-v5#isloading-and-isfetching-flags)
**Why It Happens**: `status: 'loading'` renamed to `status: 'pending'`, `isLoading` meaning changed
**Prevention**: Use `isPending` for initial load, `isLoading` for "pending AND fetching"

**Before (v4):**
```tsx
const { data, isLoading } = useQuery(...)
if (isLoading) return <div>Loading...</div>
```

**After (v5):**
```tsx
const { data, isPending, isLoading } = useQuery(...)
if (isPending) return <div>Loading...</div>
// isLoading = isPending && isFetching (fetching for first time)
```

### Issue #4: cacheTime → gcTime
**Error**: `cacheTime is not a valid option`
**Source**: [v5 Migration: gcTime](https://tanstack.com/query/latest/docs/framework/react/guides/migrating-to-v5#cachetime-has-been-replaced-by-gctime)
**Why It Happens**: Renamed to better reflect "garbage collection time"
**Prevention**: Use `gcTime` instead of `cacheTime`

**Before (v4):**
```tsx
useQuery({
  queryKey: ['todos'],
  queryFn: fetchTodos,
  cacheTime: 1000 * 60 * 60,
})
```

**After (v5):**
```tsx
useQuery({
  queryKey: ['todos'],
  queryFn: fetchTodos,
  gcTime: 1000 * 60 * 60,
})
```

### Issue #5: useSuspenseQuery + enabled
**Error**: Type error, enabled option not available
**Source**: [GitHub Discussion #6206](https://github.com/TanStack/query/discussions/6206)
**Why It Happens**: Suspense guarantees data is available, can't conditionally disable
**Prevention**: Use conditional rendering instead of `enabled` option

**Before (v4/incorrect):**
```tsx
useSuspenseQuery({
  queryKey: ['todo', id],
  queryFn: () => fetchTodo(id),
  enabled: !!id, // ❌ Not allowed
})
```

**After (v5/correct):**
```tsx
// Conditional rendering:
{id ? (
  <TodoComponent id={id} />
) : (
  <div>No ID selected</div>
)}

// Inside TodoComponent:
function TodoComponent({ id }: { id: number }) {
  const { data } = useSuspenseQuery({
    queryKey: ['todo', id],
    queryFn: () => fetchTodo(id),
    // No enabled option needed
  })
  return <div>{data.title}</div>
}
```

### Issue #6: initialPageParam Required
**Error**: `initialPageParam is required` type error
**Source**: [v5 Migration: Infinite Queries](https://tanstack.com/query/latest/docs/framework/react/guides/migrating-to-v5#new-required-initialPageParam-option)
**Why It Happens**: v4 passed `undefined` as first pageParam, v5 requires explicit value
**Prevention**: Always specify `initialPageParam` for infinite queries

**Before (v4):**
```tsx
useInfiniteQuery({
  queryKey: ['projects'],
  queryFn: ({ pageParam = 0 }) => fetchProjects(pageParam),
  getNextPageParam: (lastPage) => lastPage.nextCursor,
})
```

**After (v5):**
```tsx
useInfiniteQuery({
  queryKey: ['projects'],
  queryFn: ({ pageParam }) => fetchProjects(pageParam),
  initialPageParam: 0, // ✅ Required
  getNextPageParam: (lastPage) => lastPage.nextCursor,
})
```

### Issue #7: keepPreviousData Removed
**Error**: `keepPreviousData is not a valid option`
**Source**: [v5 Migration: placeholderData](https://tanstack.com/query/latest/docs/framework/react/guides/migrating-to-v5#removed-keeppreviousdata-in-favor-of-placeholderdata-identity-function)
**Why It Happens**: Replaced with more flexible `placeholderData` function
**Prevention**: Use `placeholderData: keepPreviousData` helper

**Before (v4):**
```tsx
useQuery({
  queryKey: ['todos', page],
  queryFn: () => fetchTodos(page),
  keepPreviousData: true,
})
```

**After (v5):**
```tsx
import { keepPreviousData } from '@tanstack/react-query'

useQuery({
  queryKey: ['todos', page],
  queryFn: () => fetchTodos(page),
  placeholderData: keepPreviousData,
})
```

### Issue #8: TypeScript Error Type Default
**Error**: Type errors with error handling
**Source**: [v5 Migration: Error Types](https://tanstack.com/query/latest/docs/framework/react/guides/migrating-to-v5#typeerror-is-now-the-default-error)
**Why It Happens**: v4 used `unknown`, v5 defaults to `Error` type
**Prevention**: If throwing non-Error types, specify error type explicitly

**Before (v4 - error was unknown):**
```tsx
const { error } = useQuery({
  queryKey: ['data'],
  queryFn: async () => {
    if (Math.random() > 0.5) throw 'custom error string'
    return data
  },
})
// error: unknown
```

**After (v5 - specify custom error type):**
```tsx
const { error } = useQuery<DataType, string>({
  queryKey: ['data'],
  queryFn: async () => {
    if (Math.random() > 0.5) throw 'custom error string'
    return data
  },
})
// error: string | null

// Or better: always throw Error objects
const { error } = useQuery({
  queryKey: ['data'],
  queryFn: async () => {
    if (Math.random() > 0.5) throw new Error('custom error')
    return data
  },
})
// error: Error | null (default)
```

### Issue #9: Streaming Server Components Hydration Error
**Error**: `Hydration failed because the initial UI does not match what was rendered on the server`
**Source**: [GitHub Issue #9642](https://github.com/TanStack/query/issues/9642)
**Affects**: v5.82.0+ with streaming SSR (void prefetch pattern)
**Why It Happens**: Race condition where `hydrate()` resolves synchronously but `query.fetch()` creates async retryer, causing isFetching/isStale mismatch between server and client
**Prevention**: Don't conditionally render based on `fetchStatus` with `useSuspenseQuery` and streaming prefetch, OR await prefetch instead of void pattern

**Before (causes hydration error):**
```tsx
// Server: void prefetch
streamingQueryClient.prefetchQuery({ queryKey: ['data'], queryFn: getData });

// Client: conditional render on fetchStatus
const { data, isFetching } = useSuspenseQuery({ queryKey: ['data'], queryFn: getData });
return <>{data && <div>{data}</div>} {isFetching && <Loading />}</>;
```

**After (workaround):**
```tsx
// Option 1: Await prefetch
await streamingQueryClient.prefetchQuery({ queryKey: ['data'], queryFn: getData });

// Option 2: Don't render based on fetchStatus with Suspense
const { data } = useSuspenseQuery({ queryKey: ['data'], queryFn: getData });
return <div>{data}</div>;  // No conditional on isFetching
```

**Status**: Known issue, being investigated by maintainers. Requires implementation of `getServerSnapshot` in useSyncExternalStore.

### Issue #10: useQuery Hydration Error with Prefetching
**Error**: Text content mismatch during hydration
**Source**: [GitHub Issue #9399](https://github.com/TanStack/query/issues/9399)
**Affects**: v5.x with server-side prefetching
**Why It Happens**: `tryResolveSync` detects resolved promises in RSC payload and extracts data synchronously during hydration, bypassing normal pending state
**Prevention**: Use `useSuspenseQuery` instead of `useQuery` for SSR, or avoid conditional rendering based on `isLoading`

**Before (causes hydration error):**
```tsx
// Server Component
const queryClient = getServerQueryClient();
await queryClient.prefetchQuery({ queryKey: ['todos'], queryFn: fetchTodos });

// Client Component
function Todos() {
  const { data, isLoading } = useQuery({ queryKey: ['todos'], queryFn: fetchTodos });
  if (isLoading) return <div>Loading...</div>;  // Server renders this
  return <div>{data.length} todos</div>;  // Client hydrates with this
}
```

**After (workaround):**
```tsx
// Use useSuspenseQuery instead
function Todos() {
  const { data } = useSuspenseQuery({ queryKey: ['todos'], queryFn: fetchTodos });
  return <div>{data.length} todos</div>;
}
```

**Status**: "At the top of my OSS list of things to fix" - maintainer Ephem (Nov 2025). Requires implementing `getServerSnapshot` in useSyncExternalStore.

### Issue #11: refetchOnMount Not Respected for Errored Queries
**Error**: Queries refetch on mount despite `refetchOnMount: false`
**Source**: [GitHub Issue #10018](https://github.com/TanStack/query/issues/10018)
**Affects**: v5.90.16+
**Why It Happens**: Errored queries with no data are always treated as stale. This is intentional to avoid permanently showing error states
**Prevention**: Use `retryOnMount: false` instead of (or in addition to) `refetchOnMount: false`

**Before (refetches despite setting):**
```tsx
const { data, error } = useQuery({
  queryKey: ['data'],
  queryFn: () => { throw new Error('Fails') },
  refetchOnMount: false,  // Ignored when query is in error state
  retry: 0,
});
// Query refetches every time component mounts
```

**After (correct):**
```tsx
const { data, error } = useQuery({
  queryKey: ['data'],
  queryFn: failingFetch,
  refetchOnMount: false,
  retryOnMount: false,  // ✅ Prevents refetch on mount for errored queries
  retry: 0,
});
```

**Status**: Documented behavior (intentional). The name `retryOnMount` is slightly misleading - it controls whether errored queries trigger a new fetch on mount, not automatic retries.

### Issue #12: Mutation Callback Signature Breaking Change (v5.89.0)
**Error**: TypeScript errors in mutation callbacks
**Source**: [GitHub Issue #9660](https://github.com/TanStack/query/issues/9660)
**Affects**: v5.89.0+
**Why It Happens**: `onMutateResult` parameter added between `variables` and `context`, changing callback signatures from 3 params to 4
**Prevention**: Update all mutation callbacks to accept 4 parameters instead of 3

**Before (v5.88 and earlier):**
```tsx
useMutation({
  mutationFn: addTodo,
  onError: (error, variables, context) => {
    // context is now onMutateResult, missing final context param
  },
  onSuccess: (data, variables, context) => {
    // Same issue
  }
});
```

**After (v5.89.0+):**
```tsx
useMutation({
  mutationFn: addTodo,
  onError: (error, variables, onMutateResult, context) => {
    // onMutateResult = return value from onMutate
    // context = mutation function context
  },
  onSuccess: (data, variables, onMutateResult, context) => {
    // Correct signature with 4 parameters
  }
});
```

**Note**: If you don't use `onMutate`, the `onMutateResult` parameter will be undefined. This breaking change was introduced in a patch version.

### Issue #13: Readonly Query Keys Break Partial Matching (v5.90.8)
**Error**: `Type 'readonly ["todos", string]' is not assignable to type '["todos", string]'`
**Source**: [GitHub Issue #9871](https://github.com/TanStack/query/issues/9871) | Fixed in [PR #9872](https://github.com/TanStack/query/pull/9872)
**Affects**: v5.90.8 only (fixed in v5.90.9)
**Why It Happens**: Partial query matching broke TypeScript types for readonly query keys (using `as const`)
**Prevention**: Upgrade to v5.90.9+ or use type assertions if stuck on v5.90.8

**Before (v5.90.8 - TypeScript error):**
```tsx
export function todoQueryKey(id?: string) {
  return id ? ['todos', id] as const : ['todos'] as const;
}
// Type: readonly ['todos', string] | readonly ['todos']

useMutation({
  mutationFn: addTodo,
  onSuccess: () => {
    queryClient.invalidateQueries({
      queryKey: todoQueryKey('123')
      // Error: readonly ['todos', string] not assignable to ['todos', string]
    });
  }
});
```

**After (v5.90.9+):**
```tsx
// Works correctly with readonly types
queryClient.invalidateQueries({
  queryKey: todoQueryKey('123')  // ✅ No type error
});
```

**Status**: Fixed in v5.90.9. Particularly affected users of code generators like `openapi-react-query` that produce readonly query keys.

### Issue #14: useMutationState Type Inference Lost
**Error**: `mutation.state.variables` typed as `unknown` instead of actual type
**Source**: [GitHub Issue #9825](https://github.com/TanStack/query/issues/9825)
**Affects**: All v5.x versions
**Why It Happens**: Fuzzy mutation key matching prevents guaranteed type inference (same issue as `queryClient.getQueryCache().find()`)
**Prevention**: Explicitly cast types in the `select` callback

**Before (type inference doesn't work):**
```tsx
const addTodo = useMutation({
  mutationKey: ['addTodo'],
  mutationFn: (todo: Todo) => api.addTodo(todo),
});

const pendingTodos = useMutationState({
  filters: { mutationKey: ['addTodo'], status: 'pending' },
  select: (mutation) => {
    return mutation.state.variables;  // Type: unknown
  },
});
```

**After (with explicit cast):**
```tsx
const pendingTodos = useMutationState({
  filters: { mutationKey: ['addTodo'], status: 'pending' },
  select: (mutation) => mutation.state.variables as Todo,
});
// Or cast the entire state:
select: (mutation) => mutation.state as MutationState<Todo, Error, Todo, unknown>
```

**Status**: Known limitation of fuzzy matching. No planned fix.

### Issue #15: Query Cancellation in StrictMode with fetchQuery
**Error**: `CancelledError` when using `fetchQuery()` with `useQuery`
**Source**: [GitHub Issue #9798](https://github.com/TanStack/query/issues/9798)
**Affects**: Development only (React StrictMode)
**Why It Happens**: StrictMode causes double mount/unmount. When `useQuery` unmounts and is the last observer, it cancels the query even if `fetchQuery()` is also running
**Prevention**: This is expected development-only behavior. Doesn't affect production

**Example:**
```tsx
async function loadData() {
  try {
    const data = await queryClient.fetchQuery({
      queryKey: ['data'],
      queryFn: fetchData,
    });
    console.log('Loaded:', data);  // Never logs in StrictMode
  } catch (error) {
    console.error('Failed:', error);  // CancelledError
  }
}

function Component() {
  const { data } = useQuery({ queryKey: ['data'], queryFn: fetchData });
  // In StrictMode, component unmounts/remounts, cancelling fetchQuery
}
```

**Workaround:**
```tsx
// Keep query observed with staleTime
const { data } = useQuery({
  queryKey: ['data'],
  queryFn: fetchData,
  staleTime: Infinity,  // Keeps query active
});
```

**Status**: Expected StrictMode behavior, not a bug. Production builds are unaffected.

### Issue #16: invalidateQueries Only Refetches Active Queries
**Error**: Inactive queries not refetching despite `invalidateQueries()` call
**Source**: [GitHub Issue #9531](https://github.com/TanStack/query/issues/9531)
**Affects**: All v5.x versions
**Why It Happens**: Documentation was misleading - `invalidateQueries()` only refetches "active" queries by default, not "all" queries
**Prevention**: Use `refetchType: 'all'` to force refetch of inactive queries

**Default behavior:**
```tsx
// Only active queries (currently being observed) will refetch
queryClient.invalidateQueries({ queryKey: ['todos'] });
```

**To refetch inactive queries:**
```tsx
queryClient.invalidateQueries({
  queryKey: ['todos'],
  refetchType: 'all'  // Refetch active AND inactive
});
```

**Status**: Documentation fixed to clarify "active" queries. This is the intended behavior.

---

## Community Tips

> **Note**: These tips come from community experts and maintainer blogs. Verify against your version.

### Tip: Query Options with Multiple Listeners

**Source**: [TkDodo's Blog - API Design Lessons](https://tkdodo.eu/blog/react-query-api-design-lessons-learned) | **Confidence**: HIGH
**Applies to**: v5.27.3+

When multiple components use the same query with different options (like `staleTime`), the "last write wins" rule applies for future fetches, but the current in-flight query uses its original options. This can cause unexpected behavior when components mount at different times.

**Example of unexpected behavior:**
```tsx
// Component A mounts first
function ComponentA() {
  const { data } = useQuery({
    queryKey: ['todos'],
    queryFn: fetchTodos,
    staleTime: 5000,  // Applied initially
  });
}

// Component B mounts while A's query is in-flight
function ComponentB() {
  const { data } = useQuery({
    queryKey: ['todos'],
    queryFn: fetchTodos,
    staleTime: 60000,  // Won't affect current fetch, only future ones
  });
}
```

**Recommended approach:**
```tsx
// Write options as functions that reference latest values
const getStaleTime = () => shouldUseLongCache ? 60000 : 5000;

useQuery({
  queryKey: ['todos'],
  queryFn: fetchTodos,
  staleTime: getStaleTime(),  // Evaluated on each render
});
```

### Tip: refetch() is NOT for Changed Parameters

**Source**: [Avoiding Common Mistakes with TanStack Query](https://www.buncolak.com/posts/avoiding-common-mistakes-with-tanstack-query-part-1/) | **Confidence**: HIGH

The `refetch()` function should ONLY be used for refreshing with the same parameters (like a manual "reload" button). For new parameters (filters, page numbers, search terms, etc.), include them in the query key instead.

**Anti-pattern:**
```tsx
// ❌ Wrong - using refetch() for different parameters
const [page, setPage] = useState(1);
const { data, refetch } = useQuery({
  queryKey: ['todos'],  // Same key for all pages
  queryFn: () => fetchTodos(page),
});

// This refetches with OLD page value, not new one
<button onClick={() => { setPage(2); refetch(); }}>Next</button>
```

**Correct pattern:**
```tsx
// ✅ Correct - include parameters in query key
const [page, setPage] = useState(1);
const { data } = useQuery({
  queryKey: ['todos', page],  // Key changes with page
  queryFn: () => fetchTodos(page),
  // Query automatically refetches when page changes
});

<button onClick={() => setPage(2)}>Next</button>  // Just update state
```

**When to use refetch():**
```tsx
// ✅ Manual refresh of same data (refresh button)
const { data, refetch } = useQuery({
  queryKey: ['todos'],
  queryFn: fetchTodos,
});

<button onClick={() => refetch()}>Refresh</button>  // Same parameters
```

---

## Key Patterns

**Dependent Queries** (Query B waits for Query A):
```tsx
const { data: posts } = useQuery({
  queryKey: ['users', userId, 'posts'],
  queryFn: () => fetchUserPosts(userId),
  enabled: !!user, // Wait for user
})
```

**Parallel Queries** (fetch multiple at once):
```tsx
const results = useQueries({
  queries: ids.map(id => ({ queryKey: ['todos', id], queryFn: () => fetchTodo(id) })),
})
```

**Prefetching** (preload on hover):
```tsx
queryClient.prefetchQuery({ queryKey: ['todo', id], queryFn: () => fetchTodo(id) })
```

**Infinite Scroll** (useInfiniteQuery):
```tsx
useInfiniteQuery({
  queryKey: ['todos', 'infinite'],
  queryFn: ({ pageParam }) => fetchTodosPage(pageParam),
  initialPageParam: 0, // Required in v5
  getNextPageParam: (lastPage) => lastPage.nextCursor,
})
```

**Query Cancellation** (auto-cancel on queryKey change):
```tsx
queryFn: async ({ signal }) => {
  const res = await fetch(`/api/todos?q=${search}`, { signal })
  return res.json()
}
```

**Data Transformation** (select):
```tsx
select: (data) => data.filter(todo => todo.completed)
```

**Avoid Request Waterfalls**: Fetch in parallel when possible (don't chain queries unless truly dependent)

---

**Official Docs**: https://tanstack.com/query/latest | **v5 Migration**: https://tanstack.com/query/latest/docs/framework/react/guides/migrating-to-v5 | **GitHub**: https://github.com/TanStack/query | **Context7**: `/websites/tanstack_query`

Overview

This skill documents practical patterns and gotchas for managing server state in React with TanStack Query (v5). It focuses on new v5 features like useMutationState, simplified optimistic updates, throwOnError, networkMode, and helpers for infinite queries. The material targets setup, v4→v5 migration fixes, and debugging SSR/hydration with streaming server components.

How this skill works

The guidance inspects common migration errors and enforces v5 conventions: object syntax for hooks, gcTime instead of cacheTime, and use of isPending for initial loading. It explains cross-component mutation tracking with useMutationState, optimistic UI using pending mutation variables, and error handling via throwOnError and QueryErrorResetBoundary. It also covers networkMode for offline/PWA behavior and typed infiniteQueryOptions with maxPages to limit memory use.

When to use it

  • When upgrading a codebase from React Query v4 to TanStack Query v5 to avoid breaking changes.
  • When implementing optimistic UI that reflects pending mutations without manual cache rollbacks.
  • When you need a global mutation tracker or UI that reflects pending saves across components.
  • When supporting offline/PWA behavior and controlling fetch strategy with networkMode.
  • When building or debugging infinite scrolling with type-safe options and memory limits.
  • When troubleshooting SSR/hydration issues with streaming server components and query state

Best practices

  • Always call hooks with object syntax: useQuery({ queryKey, queryFn, ... }) and useMutation({ mutationFn, ... }).
  • Use array-based queryKey structures for predictable cache separation: ['todos'], ['todos', id], etc.
  • Prefer isPending for initial loading state and isLoading only for pending+fetching semantics.
  • Throw errors from queryFn and use throwOnError with QueryErrorResetBoundary to integrate error boundaries.
  • Use useMutationState to show optimistic items from pending mutations instead of mutating cache directly.
  • Use queryOptions/infiniteQueryOptions factories and set gcTime (not cacheTime) to control cache lifetime

Example use cases

  • Show a global "Saving X items…" indicator by selecting pending mutations with useMutationState.
  • Render optimistic list items while POST mutations are pending using mutation variables as placeholders.
  • Fix migration TypeScript errors caused by v4 overloads by switching to object syntax everywhere.
  • Configure a PWA to read from cache when offline (networkMode: 'offlineFirst') and resume fetches when online.
  • Create a typed infinite scroll using infiniteQueryOptions and limit memory with maxPages and bi-directional params.

FAQ

Why did my onSuccess callback on useQuery stop running after upgrading?

v5 removed query callbacks (onSuccess/onError/onSettled) for useQuery. Use useEffect to react to data changes or move side effects to mutation callbacks.

What loading flag should I use for initial renders?

Use isPending for initial loading (no data yet). isLoading now means isPending && isFetching and has different semantics.