This page documents an earlier version of the API client. For the latest version, see Search an index.
search
The search query only allows for the retrieval of up to 1,000 hits.
If you need to retrieve more than 1,000 hits,
you can either use the browse
method or increase the paginationLimitedTo
parameter.
Examples
Search
Search and send an extra header
Parameters
The query used to search.
Accepts every character, and every character entered will be used in the search.
There’s a hard limit of 512 characters per query. If a search query is longer, the API will return an error.An empty query can be used to fetch all records.
A mapping of request options when using this method.
Search parameters to send with the query.
Response
If a search encounters an index that is being A/B tested,
abTestID
reports the ongoing A/B test ID.Returned only if getRankingInfo
is set to true
.If a search encounters an index that is being A/B tested,
abTestVariantID
reports the variant ID of the index used (note, this is the ID not the name).
The variant ID is the position in the array of variants (starting at 1).For example, abTestVariantID=1
is variant A (the main index), abTestVariantID=2
is variant B (the replica you chose when creating the A/B test, or the queries with the changed query parameters if the A/B test is based on query parameters).Returned only if getRankingInfo
is set to true
.The computed geo location. Warning: for legacy reasons, this parameter is a string and not an object.
Format:
${lat},${lng}
, where the latitude and longitude are expressed as decimal floating point numbers.Only returned when aroundLatLngViaIP
or aroundLatLng
is set.The automatically computed radius. For legacy reasons, this parameter is a string and not an integer.Only returned for geo queries without an explicitly specified radius (see
aroundRadius
).Whether certain properties of the search response are calculated exhaustive (exact) or approximated.List of fields:
facetsCount
: Whether the facet count is exhaustive (true
) or approximate (false
). See the related discussion.facetValues
: The value isfalse
if not all facet values are retrieved.nbHits
: Whether thenbHits
is exhaustive (true
) or approximate (false
).- When the query takes more than 50ms to be processed, the engine makes an approximation.
This can happen when using complex filters on millions of records, when typo-tolerance was not exhaustive, or when enough hits have been retrieved (for example, after the engine finds 10,000 exact matches).
nbHits
is reported as non-exhaustive whenever an approximation is made, even if the approximation didn’t, in the end, impact the exhaustivity of the query. rulesMatch
: Rules matching exhaustivity. The value isfalse
if rules were enable for this query, and could not be fully processed due a timeout. This is generally caused by the number of alternatives (such as typos) which is too large.typo
: Whether the typo search was exhaustive (true
) or approximate (false
). An approximation is done when the typo search query part takes more than 10% of the query budget (ie. 5ms by default) to be processed\n (this can happen when a lot of typo alternatives exist for the query).\n This field will not be included when typo-tolerance is entirely disabled.
See the
facetsCount
field of the exhaustive
object in the response.See the
nbHits
field of the exhaustive
object in the response.See the
typo
field of the exhaustive
object in the response.A mapping of each facet name to the corresponding facet counts.Returned only if
facets
is non-empty.Statistics for numerical facets.
${facet_name}
(object): The statistics for a given facet:min
(number
): The minimum value in the result set.max
(number
): The maximum value in the result set.avg
(number
): The average facet value in the result set.sum
(number
): The sum of all values in the result set.
maxValuesPerFacet
),
statistics are always computed on at most 1,000 distinct values (starting with the most frequent ones).
If there are more than 1,000 distinct values, statistics may therefore not be 100% accurate.Returned only if facets
is non-empty and at least one of the returned facets contains numerical values.The hits returned by the search.Hits are ordered according to the ranking
or sorting of the index being queried.Hits are made of the schemaless JSON objects that you stored in the index.However, Algolia does enrich them with a few additional fields,
such as
_highlightResult
, _snippetResult
, _rankingInfo
, and _distinctSeqID
.Example:JSON
The maximum number of hits returned per page.
See the
hitsPerPage
search parameter.Not returned if you use offset
and length
for pagination.Index name used for the query.
In the case of an A/B test, the targeted index isn’t always the index used by the query.Returned only if
getRankingInfo
is set to true
.Used to return warnings about the query.
The number of hits matched by the query.
The number of returned pages. Calculation is based on the total number of hits (
nbHits
)
divided by the number of hits per page (hitsPerPage
), rounded up to the nearest integer.Not returned if you use offset
and length
for pagination.The number of hits selected and sorted by the relevant sort algorithm.Only returned when searching on a virtual replica.
Index of the current page (zero-based). See the
page
search parameter.Not returned if you use offset
/length
for pagination.A url-encoded string of all search parameters.
The query string that will be searched, after normalization.
Normalization includes removing stop words (if
removeStopWords
is enabled),
and transforming portions of the query string into phrase queries (see advancedSyntax
).Returned only if getRankingInfo
is set to true
.Time the server took to process the search request, in milliseconds.
This doesn’t include network time.
List of processing steps and their times, in milliseconds. You can use this list to investigate performance issues.
This parameter is experimental.
Fields may change without further notice and may be different for different Algolia servers.
The list of steps is not exhaustive.
A markup text indicating which parts of the original query have been removed
in order to retrieve a non-empty result set.The removed parts are surrounded by
<em>
tags.Only returned when removeWordsIfNoResults
is set to lastWords
or firstWords
.Unique identifier of the search query,
for sending events to the Insights API.Returned only if
clickAnalytics
is true
.Time the server took to process the complete request, in milliseconds.
This does not include network time.
Actual host name of the server that processed the request.
The DNS supports automatic failover and load balancing,
so this may differ from the host name used in the request.Returned only if
getRankingInfo
is set to true
.Use the
facetsCount
field of the exhaustive
object in the response.Returned only if getRankingInfo
is set to true
.Please use
exhaustiveHitsCount
instead.Returned only if getRankingInfo
is set to true
.Response as JSON
This section shows the JSON response returned by the API. Each API client wraps this response in language-specific objects, so the structure may vary. To view the response, use thegetLogs
method.
Don’t rely on the order of properties—JSON objects don’t preserve key order.
JSON