pagination is a UI pattern for navigation across (typically chronologically) sequential pages that show one or more posts such as permalink post pages, archives, search results, and lists of tagged posts.
Date Based Pagination
- Aaron Parecki uses date-based pagination on aaronparecki.com. This means each page's URL is based off of the published date of the top post on the page, so this URL will always show the same set of posts, even as new posts are added. The pages in this way become permalinks, rather than dynamically changing.
- Kartik Prabhu uses date-based pagination on kartikprabhu.com. Articles are paginated by year whereas more-frequent Notes by month.
Slack uses a cursor-based pagination mechanism. An API call that returns a subset of a list of items will also return a cursor string which can be used to retrieve the next page of results. The cursor string is opaque to the client, and only has significance to the server.
- Paginated methods accept limit and cursor parameters.
- Paginated requests include a top-level response_metadata object including a next_cursor when additional results can be retrieved.
- Present the cursor parameter on subsequent requests to navigate the collection.
- An empty, null, or non-existent next_cursor indicates no further results.
Facebook supports multiple pagination mechanisms, but encourages the use of cursor-based pagination. Facebook returns two cursors, "after" and "before" to traverse both directions in a list.
The "before" cursor points to the start of the page of data. The "after" cursor points to the end of the page of data. To request the previous or next page of data, you can pass these strings in the "before" or "after" query string parameter of the API request.
MediaWiki's web interface for contributions, recent changes, etc. use query params with:
- offset=YYYYMMDDHHMMSS and
MediaWiki's API for recent changes uses start/end timestamp-based parameters, where each parameter name starts with the prefix for the method group, e.g. "rcend" for "recent changes end". There is an additional parameter to indicate whether to return results sorted by
- rcdir = older or newer
The Mediawiki API uses cursor-style pagination:
Google Web Search
Google's web search uses query params for paginating search results with:
Most of Twitter's APIs for returning tweets use a paging mechanism built around their use of always-increasing numeric tweet IDs. Their IDs are based on the timestamp rather than being sequential, but are guaranteed to be sorted by the time the tweet was created.
Paging in the Twitter API involves using either the
max_id parameters to return tweets newer or older than the specified tweet respectively.
Twitter also has a section of their documentation about their cursor-based API.
API responses include
previous_cursor, and will return "0" if there are no more results.
Twitter Ads API
The Twitter Ads API uses cursor-based pagination using the
count parameters. Traversing
If a response has more than
countentities, these endpoints will now return the first
countentities and a
next_cursorkey in the response JSON.
If less than
countentities are returned in the current page of the result set, the
next_cursorvalue will be null.
IndieWebCamp sessions that discussed pagination: