Interface: Query<TableInfo>

server.Query

The Query interface allows functions to read values out of the database.

If you only need to load an object by ID, use db.get(id) instead.

Executing a query consists of calling

1. (Optional) order to define the order
2. (Optional) filter to refine the results
3. A consumer method to obtain the results

Queries are lazily evaluated. No work is done until iteration begins, so constructing and extending a query is free. The query is executed incrementally as the results are iterated over, so early terminating also reduces the cost of the query.

Example

// Use .withIndex() for efficient queries (preferred over .filter()):
const messages = await ctx.db
  .query("messages")
  .withIndex("by_channel", (q) => q.eq("channelId", channelId))
  .order("desc")
  .take(10);

// Async iteration for processing large result sets:
for await (const task of ctx.db.query("tasks")) {
  // Process each task without loading all into memory
}

// Get a single unique result (throws if multiple match):
const user = await ctx.db
  .query("users")
  .withIndex("by_email", (q) => q.eq("email", email))
  .unique();

Common mistake: .collect() loads all matching documents into memory. If the result set can grow unbounded as your database grows, this will eventually cause problems. Prefer .first(), .unique(), .take(n), or pagination instead. Only use .collect() on queries with a tightly bounded result set (e.g., items belonging to a single user with a known small limit).

Ordering
order("asc")	Define the order of query results.
Filtering
filter(...)	Filter the query results to only the values that match some condition.
Consuming	Execute a query and return results in different ways.
[Symbol.asyncIterator]()	The query's results can be iterated over using a for await..of loop.
collect()	Return all of the results as an array.
take(n: number)	Return the first n results as an array.
first()	Return the first result.
unique()	Return the only result, and throw if there is more than one result.

To learn more about how to write queries, see Querying the Database.

Type parameters

Name	Type
TableInfo	extends GenericTableInfo

Hierarchy

• OrderedQuery<TableInfo>

  ↳ Query

  ↳↳ QueryInitializer

Methods

[asyncIterator]

▸ [asyncIterator](): AsyncIterator<DocumentByInfo<TableInfo>, any, undefined>

Returns

AsyncIterator<DocumentByInfo<TableInfo>, any, undefined>

Inherited from

OrderedQuery.[asyncIterator]

Defined in

../../common/temp/node_modules/.pnpm/typescript@5.0.4/node_modules/typescript/lib/lib.es2018.asynciterable.d.ts:38

---

order

▸ order(order): OrderedQuery<TableInfo>

Define the order of the query output.

Use "asc" for an ascending order and "desc" for a descending order. If not specified, the order defaults to ascending.

Parameters

Name	Type	Description
order	"asc" | "desc"	The order to return results in.

Returns

OrderedQuery<TableInfo>

Defined in

server/query.ts:174

---

filter

▸ filter(predicate): Query<TableInfo>

Filter the query output, returning only the values for which predicate evaluates to true.

Important: Prefer using .withIndex() over .filter() whenever possible. Filters scan all documents matched so far and discard non-matches, while indexes efficiently skip non-matching documents. Define an index in your schema for fields you filter on frequently.

Parameters

Name	Type	Description
predicate	(q: FilterBuilder<TableInfo>) => ExpressionOrValue<boolean>	An Expression constructed with the supplied FilterBuilder that specifies which documents to keep.

Returns

Query<TableInfo>

• A new OrderedQuery with the given filter predicate applied.

Inherited from

OrderedQuery.filter

Defined in

server/query.ts:195

---

paginate

▸ paginate(paginationOpts): Promise<PaginationResult<DocumentByInfo<TableInfo>>>

Load a page of n results and obtain a Cursor for loading more.

Note: If this is called from a reactive query function the number of results may not match paginationOpts.numItems!

paginationOpts.numItems is only an initial value. After the first invocation, paginate will return all items in the original query range. This ensures that all pages will remain adjacent and non-overlapping.

Parameters

Name	Type	Description
paginationOpts	PaginationOptions	A PaginationOptions object containing the number of items to load and the cursor to start at.

Returns

Promise<PaginationResult<DocumentByInfo<TableInfo>>>

A PaginationResult containing the page of results and a cursor to continue paginating.

Inherited from

OrderedQuery.paginate

Defined in

server/query.ts:224

---

collect

▸ collect(): Promise<DocumentByInfo<TableInfo>[]>

Execute the query and return all of the results as an array.

Warning: This loads every matching document into memory. If the result set can grow unbounded as your database grows, .collect() will eventually cause performance problems or hit limits. Only use .collect() when the result set is tightly bounded (e.g., a known small number of items).

Prefer .first(), .unique(), .take(n), or .paginate() when the result set may be large or unbounded. For processing many results without loading all into memory, use the Query as an AsyncIterable with for await...of.

Returns

Promise<DocumentByInfo<TableInfo>[]>

• An array of all of the query's results.

Inherited from

OrderedQuery.collect

Defined in

server/query.ts:243

---

take

▸ take(n): Promise<DocumentByInfo<TableInfo>[]>

Execute the query and return the first n results.

Parameters

Name	Type	Description
n	number	The number of items to take.

Returns

Promise<DocumentByInfo<TableInfo>[]>

• An array of the first n results of the query (or less if the query doesn't have n results).

Inherited from

OrderedQuery.take

Defined in

server/query.ts:252

---

first

▸ first(): Promise<null | DocumentByInfo<TableInfo>>

Execute the query and return the first result if there is one.

Returns

Promise<null | DocumentByInfo<TableInfo>>

• The first value of the query or null if the query returned no results.

Inherited from

OrderedQuery.first

Defined in

server/query.ts:259

---

unique

▸ unique(): Promise<null | DocumentByInfo<TableInfo>>

Execute the query and return the singular result if there is one.

Use this when you expect exactly zero or one result, for example when querying by a unique field. If the query matches more than one document, this will throw an error.

Example

const user = await ctx.db
  .query("users")
  .withIndex("by_email", (q) => q.eq("email", "alice@example.com"))
  .unique();

Throws

Will throw an error if the query returns more than one result.

Returns

Promise<null | DocumentByInfo<TableInfo>>

• The single result returned from the query or null if none exists.

Inherited from

OrderedQuery.unique

Defined in

server/query.ts:279