Infinite scroll

2 min read

The last row of a table asks for the next page as soon as it comes into view. The answer adds more rows after it, and the new last row asks for the page after that. When no pages are left, the answer ends without a trigger.

Try it

Live demo

NameEmail
Marina Duarte[email protected]
Tomás Ribeiro[email protected]
Helena Prado[email protected]
Caio Nogueira[email protected]
Beatriz Lemos[email protected]
Rafael Quintana[email protected]
Inês Carvalho[email protected]
Joaquim Teles[email protected]
Larissa Moura[email protected]

Real requests to this site. What you change is kept in a cookie of your own for an hour and is invisible to every other visitor.

Scroll inside the box. Each page holds ten rows, and the last page has no trigger, so the requests stop.

The page

html
<table>
  <thead><tr><th>Name</th><th>Email</th></tr></thead>
  <tbody>
    <tr><td>Marina Duarte</td><td>[email protected]</td></tr>
    <!-- … eight more rows … -->
    <tr ypy-get="/contacts?page=2" ypy-trigger="revealed" ypy-swap="afterend">
      <td>Caio Nogueira</td><td>[email protected]</td>
    </tr>
  </tbody>
</table>

What the server answers

GET /contacts?page=2 answers 200 with the rows of that page. The last of them carries the next trigger:

html
<tr><td>Beatriz Lemos</td><td>[email protected]</td></tr>
<!-- … eight more rows … -->
<tr ypy-get="/contacts?page=3" ypy-trigger="revealed" ypy-swap="afterend">
  <td>Rafael Quintana</td><td>[email protected]</td>
</tr>
Case Status Body
A page with more after it 200 The rows; the last row has ypy-get for the next page
The last page 200 The rows, none of them with ypy-get
A page past the end 200 Empty

Every answer is a fragment of <tr> elements with Vary: YPY-Request.

How it works

  • ypy-trigger="revealed" fires once, when the row scrolls into the viewport. The observer then disconnects, so a fast scroll cannot ask for the same page twice.

  • ypy-swap="afterend" puts the answer after the row that asked. The old last row stays where it is and keeps its attributes, but nothing watches it any more.

  • The new last row is the one that carries ypy-get, so the chain continues for as long as the server keeps handing out a next address.

  • The server decides when to stop: a last page whose rows have no ypy-get ends the chain.
  • Rows are parsed inside a <template>, so <tr> elements arrive intact.

Variations

  • Start loading before the visitor reaches the end: put the trigger on a row some lines above the last one and use ypy-trigger="intersect once", or intersect root:#scroller threshold:0 when the rows scroll inside a box with that id.

  • Visitors without JavaScript see only the first page. Close the table with a normal link to the next page, so the content stays reachable.

  • The ypy-list extension keeps a growing list bounded with limit, so a long session does not hold thousands of rows.

See also