Docs · Sources
Pagination
Many APIs return records in pages. Set Pagination so Rowvela requests every page in one run. After Inspect source, Rowvela usually recognises the pattern and suggests the right setting.
Modes
| Pagination | The API… |
|---|---|
| Single response | returns all records at once. |
| Page number | takes a page parameter such as page=2. |
| Record offset | takes the number of records to skip, such as offset=100. |
| Cursor | returns a token for the next page, in a response field or as the last item’s ID. |
| Next page URL in the response | returns the full URL of the next page, for example links.next. |
| Link header | sends the next page in an HTTP Link header. |
Set the Parameter name, Starting value and Page size the API expects. Send pagination in chooses URL parameters or the JSON body, request headers, or a {page} placeholder in the URL path.
When pagination ends
- Stop at an empty page is the safe default: a shorter page does not end the import early.
- API always returns full pages · stop at a shorter page saves one request when the API is reliable about page sizes.
- A cursor import can also stop on a Has-more flag such as
has_more. - If a page or cursor repeats, the run fails instead of looping or duplicating rows. If the page limit is reached while the source still has data, the run fails instead of importing only part of it.
Next-page URLs must stay on the source host, so your credentials are never sent to another server.