Pagination
On this page
Pagination gives you full control over how you retrieve query results. You can implement standard paging or retrieve specific record subsets that ignore page boundaries.
You can set pagination defaults at indexing time and override them at query time.
Pagination at index time
At indexing time you can set up 2 defaults:
- hitsPerPage - the number of hits returned for every page
- paginationLimitedTo - controls the overall number of hits you can potentially retrieve for a given query
Pagination at query time
Response details
In order to help paginate results, Algolia provides information about the current state of the pagination in the JSON response.
1
2
3
4
5
6
7
8
{
[...],
"page": 0,
"nbHits": 40,
"nbPages": 2,
"hitsPerPage": 20,
"exhaustiveNbHits": true
}
page
returns the current page (note: this value is zero-based)hitsPerPage
returns the maximum number of hits returned for each pagenbPages
returns the number of pages available for the current querynbHits
returns the number of hits that the search query matchedexhaustiveNbHits
returns a boolean indicating if the nbHits count was exhaustive or approximate
Exhaustivity
If possible, Algolia will return an accurate number of total hits. However, there are some cases where performance needs to be favored over exhaustivity. If a query returns a huge number of results, the engine will approximate the hits count to avoid having to scan the full results set. This approximation has been put in place to protect other search and indexing operations. As an alternative, you can leverage the boolean exhaustiveNbHits to either hide or tweak the display of the hits count in this case. You can also use the frequent occurrence of this value to consider fine-tuning your data and index settings to improve performance.
Retrieving specific pages
At query time, you can pass in the page parameter to access directly the nth page of results. The hitsPerPage parameter provides a way to set the default number of hits returned for each page. By default, its value is 20; however, this can be overridden either as an index setting or as a query parameter.
1
2
3
4
5
index.search({
query: 'query',
page: 2,
hitsPerPage: 5
});
Pagination Limitations
1000 hit limit
By default, Algolia limits the maximum number of hits that can be retrieved via a query to 1000. This restriction is in place to not only guarantee optimal performance, but also to ensure your data cannot be easily extracted. For the vast majority of use cases, the default of 1000 hits is more than enough.
Overriding the Default Limits
In the event of a specific use case that requires being able to retrieve a higher number of hits (for example, powering an e-commerce category page or developing an infinite scroll type of experience), it is possible to raise the default pagination limit via the paginationLimitedTo index setting.
1
2
3
index.setSettings({
paginationLimitedTo: 5000
});
Increasing the pagination limit may impact performance. We highly recommend setting this number only as large as is actually necessary.
Paging beyond the limit
If you send a request for a page that does not exist, or is out-of-range (i.e. when page >= nbPages
), we do not return an error. Instead, we return 0 results.
Retrieving a subset of records (with offset and length)
Leveraging page
and hitsPerPage
is best practice for adding pagination functionality. However, there are some cases where you will want to conditionally insert items into the results (for example, ads) without impacting the pagination. In this situation, while you can still use page
and hitsPerPage
, you can also use offset
and length
.
Note that with our InstantSearch, we have implemented all pagination use cases with page
and hitsPerPage
.
The parameters offset and length provide an alternative to paging by letting you specify the records to retrieve, not the page. You may want to do this for a variety of reasons, one of which is mentioned above - to mix ads within a small subset of results.
offset
lets you specify the starting hit (or record) and length
sets the number of records returned.
For example, if you have 100 records in your result set, which are broken down into 10 pages (hitsPerPage=10
), you can ignore the pages and retrieve record subsets, as follows (offset is zero-based):
- records 50 to 80 (
offset=49
andlength = 30
). - records 21 to 25 (
offset=20
andlength = 5
).