Pagination
The Lumar API uses cursor-based pagination for all connection fields. Every connection exposes the pageInfo object:
type PageInfo {
hasNextPage: Boolean!
hasPreviousPage: Boolean!
startCursor: String
endCursor: String
}
Where hasNextPage and hasPreviousPage will tell us whether there is a next or previous page or we hit the last or first one respectively.
On the other hand, startCursor and endCursor will give us cursor values for the start and end of the collection retrieved. These can be used as the after or before parameter in subsequent queries.
Pagination parameters
| Parameter | Type | Description |
|---|---|---|
first | Int | Return the first N items from the start of the set. Use with after. |
after | String | Cursor to start after (from endCursor). |
last | Int | Return the last N items from the end of the set. Use with before. |
before | String | Cursor to end before (from startCursor). |
Use first / after for forward pagination and last / before for backward pagination.
Forward pagination example
We run the following query:
query firstPage {
me {
accounts(first: 1) {
nodes {
projects(first: 3) {
pageInfo {
hasNextPage
endCursor
}
nodes {
name
}
}
}
}
}
}
We can then get the second page with the query:
query secondPage {
me {
accounts(first: 1) {
nodes {
projects(first: 3, after: "Mw") {
pageInfo {
hasNextPage
endCursor
}
nodes {
name
}
}
}
}
}
}
Backward pagination
You can also paginate backward from the end of a collection using last and before:
query LastPageProjects {
me {
accounts(first: 1) {
nodes {
projects(last: 3) {
pageInfo {
hasPreviousPage
startCursor
}
nodes {
name
}
}
}
}
}
}
To get the previous page, pass the startCursor value as the before parameter:
query PreviousPageProjects {
me {
accounts(first: 1) {
nodes {
projects(last: 3, before: "OA") {
pageInfo {
hasPreviousPage
startCursor
}
nodes {
name
}
}
}
}
}
}
Pagination loop in TypeScript
For programmatic access, loop through all pages until hasNextPage is false:
async function fetchAllPages<T>(
queryFn: (cursor: string | null) => Promise<{
nodes: T[];
pageInfo: { hasNextPage: boolean; endCursor: string | null };
}>,
): Promise<T[]> {
const allNodes: T[] = [];
let cursor: string | null = null;
let hasNextPage = true;
while (hasNextPage) {
const page = await queryFn(cursor);
allNodes.push(...page.nodes);
hasNextPage = page.pageInfo.hasNextPage;
cursor = page.pageInfo.endCursor;
}
return allNodes;
}
Tips for large datasets
- Use a reasonable page size (50--500) depending on the number of fields requested.
- For very large exports (tens of thousands of records), consider using report downloads instead of paginating through the API.
- Include
totalCountin your first request to estimate the total number of pages. - Avoid combining
firstandlastin the same request -- use one direction at a time.