Sphinx::Search(3) Sphinx search engine API Perl client


Please note that you *MUST* install a version which is compatible with your version of Sphinx.

Use version 0.29 for Sphinx-2.2.8-release or later (or use DBI instead)

Use version 0.28 for Sphinx-2.0.8-release or later

Use version 0.27.2 for Sphinx-2.0.3-release (svn-r3043)

Use version 0.26.1 for Sphinx-2.0.1-beta (svn-r2792)

Use version 0.25_03 for Sphinx svn-r2575

Use version 0.24.1 for Sphinx-1.10-beta (svn-r2420)

Use version 0.23_02 for Sphinx svn-r2269 (experimental)

Use version 0.22 for Sphinx 0.9.9-rc2 and later (Please read the Compatibility Note under SetEncoders regarding encoding changes)

Use version 0.15 for Sphinx 0.9.9-svn-r1674

Use version 0.12 for Sphinx 0.9.8

Use version 0.11 for Sphinx 0.9.8-rc1

Use version 0.10 for Sphinx 0.9.8-svn-r1112

Use version 0.09 for Sphinx 0.9.8-svn-r985

Use version 0.08 for Sphinx 0.9.8-svn-r871

Use version 0.06 for Sphinx 0.9.8-svn-r820

Use version 0.05 for Sphinx 0.9.8-cvs-20070907

Use version 0.02 for Sphinx 0.9.8-cvs-20070818


use Sphinx::Search;
$sph = Sphinx::Search->new();
# Standard API query
$results = $sph->SetSortMode(SPH_SORT_RELEVANCE)
->Query("search terms");
# SphinxQL query
$results = $sph->SphinxQL("SELECT * FROM myindex WHERE MATCH('search terms')");


This is the Perl API client for the Sphinx open-source SQL full-text indexing search engine, <http://www.sphinxsearch.com>.

Since 0.9.9, Sphinx supports a native MySQL-protocol client, i.e. DBI with DBD::mysql. That is, you can configure the server to have a mysql41 listening port and then simply do

  my $dbh = DBI->connect('dbi:mysql:host=;port=9306;mysql_enable_utf8=1') or die "Failed to connect via DBI";
  my $sth = $dbh->prepare_cached("SELECT * FROM myindex WHERE MATCH('search terms')");
  while (my $row = $sth->fetchrow_arrayref) {
      ... # Collect results

The DBI client turns out to be significantly (about 5x) faster than this pure-Perl API. You should probably be using that instead.

This module also supports SphinxQL queries, with the small advantage that you can use either the standard API or the SphinxQL API over the one port (i.e. the searchd server does not need to be configured with a mysql41 listening port).

Given that the DBI client has several advantages over this API, future updates of this module are unlikely.



    $sph = Sphinx::Search->new;
    $sph = Sphinx::Search->new(\%options);

Create a new Sphinx::Search instance.


Specify an optional logger instance. This can be any class that provides error, warn, info, and debug methods (e.g. see Log::Log4perl). Logging is disabled if no logger instance is provided.
Debug flag. If set (and a logger instance is specified), debugging messages will be generated.



    $error = $sph->GetLastError;

Get last error message (string)


    $warning = $sph->GetLastWarning;

Get last warning message (string)


Check connection error flag (to differentiate between network connection errors and bad responses). Returns true value on connection error.


    $sph->SetEncoders(\&encode_function, \&decode_function)

COMPATIBILITY NOTE: SetEncoders() was introduced in version 0.17. Prior to that, all strings were considered to be sequences of bytes which may have led to issues with multi-byte characters. If you were previously encoding/decoding strings external to Sphinx::Search, you will need to disable encoding/decoding by setting Sphinx::Search to use raw values as explained below (or modify your code and let Sphinx::Search do the recoding).

Set the string encoder/decoder functions for transferring strings between perl and Sphinx. The encoder should take the perl internal representation and convert to the bytestream that searchd expects, and the decoder should take the bytestream returned by searchd and convert to perl format.

The searchd format will depend on the 'charset_type' index setting in the Sphinx configuration file.

The coders default to encode_utf8 and decode_utf8 respectively, which are compatible with the 'utf8' charset_type.

If either the encoder or decoder functions are left undefined in the call to SetEncoders, they return to their default values.

If you wish to send raw values (no encoding/decoding), supply a function that simply returns its argument, e.g.

    $sph->SetEncoders( sub { shift }, sub { shift });

Returns $sph.


    $sph->SetServer($host, $port);
    $sph->SetServer($path, $port);

In the first form, sets the host (string) and port (integer) details for the searchd server using a network (INET) socket (default is localhost:9312).

In the second form, where $path is a local filesystem path (optionally prefixed by 'unix://'), sets the client to access the searchd server via a local (UNIX domain) socket at the specified path.

Returns $sph.



Set server connection timeout (in seconds).

Returns $sph.



Set server connection retries (in case of connection fail).

Returns $sph.


    $sph->SetLimits($offset, $limit);
    $sph->SetLimits($offset, $limit, $max);

Set match offset/limits, and optionally the max number of matches to return.

Returns $sph.



Set maximum query time, in milliseconds, per index.

The value may not be negative; 0 means ``do not limit''.

Returns $sph.




Set match mode, which may be one of:


    Match all words


    Match any words


    Exact phrase match


    Boolean match, using AND (&), OR (|), NOT (!,-) and parenthetic grouping.


    Extended match, which includes the Boolean syntax plus field, phrase and proximity operators.

Returns $sph.


    $sph->SetRankingMode(SPH_RANK_BM25, $rank_exp);

Set ranking mode, which may be one of:


    Default mode, phrase proximity major factor and BM25 minor one


    Statistical mode, BM25 ranking only (faster but worse quality)


    No ranking, all matches get a weight of 1


    Simple word-count weighting, rank is a weighted sum of per-field keyword occurence counts


    Returns rank as it was computed in SPH_MATCH_ANY mode earlier, and is internally used to emulate SPH_MATCH_ANY queries.


    Returns a 32-bit mask with N-th bit corresponding to N-th fulltext field, numbering from 0. The bit will only be set when the respective field has any keyword occurences satisfiying the query.


    SPH_RANK_SPH04 is generally based on the default SPH_RANK_PROXIMITY_BM25 ranker, but additionally boosts the matches when they occur in the very beginning or the very end of a text field.


    Allows the ranking formula to be specified at run time. It exposes a number of internal text factors and lets you define how the final weight should be computed from those factors. $rank_exp should be set to the ranking expression string, e.g. to emulate SPH_RANK_PROXIMITY_BM25, use ``sum(lcs*user_weight)*1000+bm25''.

Returns $sph.


    $sph->SetSortMode($mode, $sortby);

Set sort mode, which may be any of:

SPH_SORT_RELEVANCE - sort by relevance
Sort by attribute descending/ascending. $sortby specifies the sorting attribute.
Sort by time segments (last hour/day/week/month) in descending order, and then by relevance in descending order. $sortby specifies the time attribute.
Sort by SQL-like syntax. $sortby is the sorting specification.

Returns $sph.



    $sph->SetWeights([ 1, 2, 3, 4]);

This method is deprecated. Use SetFieldWeights instead.

Set per-field (integer) weights. The ordering of the weights correspond to the ordering of fields as indexed.

Returns $sph.



Set per-field (integer) weights by field name. The weights hash provides field name to weight mappings.

Takes precedence over SetWeights.

Unknown names will be silently ignored. Missing fields will be given a weight of 1.

Returns $sph.



Set per-index (integer) weights. The weights hash is a mapping of index name to integer weight.

Returns $sph.


    $sph->SetIDRange($min, $max);

Set IDs range only match those records where document ID is between $min and $max (including $min and $max)

Returns $sph.


    $sph->SetFilter($attr, \@values);
    $sph->SetFilter($attr, \@values, $exclude);

Sets the results to be filtered on the given attribute. Only results which have attributes matching the given values will be returned. (Attribute values must be integers).

This may be called multiple times with different attributes to select on multiple attributes.

If 'exclude' is set, excludes results that match the filter.

Returns $sph.


    $sph->SetFilterString($attr, $value)
    $sph->SetFilterString($attr, $value, $exclude)

Adds new string value filter. Only those documents where $attr column value matches the string value from $value will be matched (or rejected, if $exclude is true).


    $sph->SetFilterRange($attr, $min, $max);
    $sph->SetFilterRange($attr, $min, $max, $exclude);

Sets the results to be filtered on a range of values for the given attribute. Only those records where $attr column value is between $min and $max (including $min and $max) will be returned.

If 'exclude' is set, excludes results that fall within the given range.

Returns $sph.


    $sph->SetFilterFloatRange($attr, $min, $max, $exclude);

Same as SetFilterRange, but allows floating point values.

Returns $sph.


    $sph->SetGeoAnchor($attrlat, $attrlong, $lat, $long);

Setup anchor point for using geosphere distance calculations in filters and sorting. Distance will be computed with respect to this point

$attrlat is the name of latitude attribute
$attrlong is the name of longitude attribute
$lat is anchor point latitude, in radians
$long is anchor point longitude, in radians

Returns $sph.


    $sph->SetGroupBy($attr, $func);
    $sph->SetGroupBy($attr, $func, $groupsort);

Sets attribute and function of results grouping.

In grouping mode, all matches are assigned to different groups based on grouping function value. Each group keeps track of the total match count, and the best match (in this group) according to current sorting function. The final result set contains one best match per group, with grouping function value and matches count attached.

$attr is any valid attribute. Use ResetGroupBy to disable grouping.

$func is one of:


    Group by day (assumes timestamp type attribute of form YYYYMMDD)


    Group by week (assumes timestamp type attribute of form YYYYNNN)


    Group by month (assumes timestamp type attribute of form YYYYMM)


    Group by year (assumes timestamp type attribute of form YYYY)


    Group by attribute value


    Group by two attributes, being the given attribute and the attribute that immediately follows it in the sequence of indexed attributes. The specified attribute may therefore not be the last of the indexed attributes.

Groups in the set of results can be sorted by any SQL-like sorting clause, including both document attributes and the following special internal Sphinx attributes:

@id - document ID;
@weight, @rank, @relevance - match weight;
@group - group by function value;
@count - number of matches in group.

The default mode is to sort by groupby value in descending order, ie. by ``@group desc''.

In the results set, ``total_found'' contains the total amount of matching groups over the whole index.

WARNING: grouping is done in fixed memory and thus its results are only approximate; so there might be more groups reported in total_found than actually present. @count might also be underestimated.

For example, if sorting by relevance and grouping by a ``published'' attribute with SPH_GROUPBY_DAY function, then the result set will contain only the most relevant match for each day when there were any matches published, with day number and per-day match count attached, and sorted by day number in descending order (ie. recent days first).



Set count-distinct attribute for group-by queries


    $sph->SetRetries($count, $delay);

Set distributed retries count and delay


    $sph->SetOverride($attrname, $attrtype, $values);
 Set attribute values override. There can be only one override per attribute.
 $values must be a hash that maps document IDs to attribute values



Set select list (attributes or expressions). SQL-like syntax.


    $sph->SetQueryFlag($flag_name, $flag_value);


  $sph->SetOuterSelect($orderby, $offset, $limit)



Clear all filters.



Clear all group-by settings (for multi-queries)


Clear all attribute value overrides (for multi-queries)


Clear all query flags.


Clear all outer select settings.


    $results = $sph->Query($query, $index);

Connect to searchd server and run given search query.

query is query string
index is index name to query, default is "*" which means to query all indexes. Use a space or comma separated list to search multiple indexes.

Returns undef on failure

Returns hash which has the following keys on success:

Array containing hashes with found documents ( ``doc'', ``weight'', ``group'', ``stamp'' )
Total amount of matches retrieved (upto SPH_MAX_MATCHES, see sphinx.h)
Total amount of matching documents in index
Search time
Hash which maps query terms (stemmed!) to ( ``docs'', ``hits'' ) hash

Returns the results array on success, undef on error.


   $sph->AddQuery($query, $index);

Add a query to a batch request.

Batch queries enable searchd to perform internal optimizations, if possible; and reduce network connection overheads in all cases.

For instance, running exactly the same query with different groupby settings will enable searched to perform expensive full-text search and ranking operation only once, but compute multiple groupby results from its output.

Parameters are exactly the same as in Query() call.

Returns corresponding index to the results array returned by RunQueries() call.



Run batch of queries, as added by AddQuery.

Returns undef on network IO failure.

Returns an array of result sets on success.

Each result set in the returned array is a hash which contains the same keys as the hash returned by Query, plus:

  • error

    Errors, if any, for this query.

  • warning

    Any warnings associated with the query.


  my $results = $sph->SphinxQL($sphinxql_query);

This is an alternative implementation of the SphinxQL API to the DBI option. Frankly, it was an experiment, and the DBI driver proved to have much better performance. Whilst this may be useful to some, in general if you are considering using this method then you should probably look at connecting directly via DBI instead.

Results are return in a hash containing an array of 'columns' and 'rows' and possibly a warning count. If a server-side error occurs, the hash contains the 'error' field. If a communication error occurs, the return value will be undefined. In either error case, GetLastError will return the error.


    $excerpts = $sph->BuildExcerpts($docs, $index, $words, $opts)

Generate document excerpts for the specified documents.

An array reference of strings which represent the document contents
A string specifiying the index whose settings will be used for stemming, lexing and case folding
A string which contains the words to highlight
A hash which contains additional optional highlighting parameters:
before_match - a string to insert before a set of matching words, default is "<b>"
after_match - a string to insert after a set of matching words, default is "<b>"
chunk_separator - a string to insert between excerpts chunks, default is " ... "
limit - max excerpt size in symbols (codepoints), default is 256
limit_passages - Limits the maximum number of passages that can be included into the snippet. Integer, default is 0 (no limit).
limit_words - Limits the maximum number of keywords that can be included into the snippet. Integer, default is 0 (no limit).
around - how many words to highlight around each match, default is 5
exact_phrase - whether to highlight exact phrase matches only, default is false
single_passage - whether to extract single best passage only, default is false
weight_order - Whether to sort the extracted passages in order of relevance (decreasing weight), or in order of appearance in the document (increasing position). Boolean, default is false.
query_mode - Whether to handle $words as a query in extended syntax, or as a bag of words (default behavior). For instance, in query mode ("one two" | "three four") will only highlight and include those occurrences "one two" or "three four" when the two words from each pair are adjacent to each other. In default mode, any single occurrence of "one", "two", "three", or "four" would be highlighted. Boolean, default is false.
force_all_words - Ignores the snippet length limit until it includes all the keywords. Boolean, default is false.
start_passage_id - Specifies the starting value of %PASSAGE_ID% macro (that gets detected and expanded in before_match, after_match strings). Integer, default is 1.
load_files - Whether to handle $docs as data to extract snippets from (default behavior), or to treat it as file names, and load data from specified files on the server side. Boolean, default is false.
html_strip_mode - HTML stripping mode setting. Defaults to "index", which means that index settings will be used. The other values are "none" and "strip", that forcibly skip or apply stripping irregardless of index settings; and "retain", that retains HTML markup and protects it from highlighting. The "retain" mode can only be used when highlighting full documents and thus requires that no snippet size limits are set. String, allowed values are "none", "strip", "index", and "retain".
allow_empty - Allows empty string to be returned as highlighting result when a snippet could not be generated (no keywords match, or no passages fit the limit). By default, the beginning of original text would be returned instead of an empty string. Boolean, default is false.

Returns undef on failure.

Returns an array ref of string excerpts on success.


    $results = $sph->BuildKeywords($query, $index, $hits)

Generate keyword list for a given query Returns undef on failure, Returns an array of hashes, where each hash describes a word in the query with the following keys:

  • tokenized

    Tokenised term from query

  • normalized

    Normalised term from query

  • docs

    Number of docs in which word was found (if $hits is true)

  • hits

    Number of occurrences of word (if $hits is true)


    $escaped = $sph->EscapeString('abcde!@#$%')

Inserts backslash before all non-word characters in the given string.


    $sph->UpdateAttributes($index, \@attrs, \%values);
    $sph->UpdateAttributes($index, \@attrs, \%values, $mva);
    $sph->UpdateAttributes($index, \@attrs, \%values, $mva, $ignorenonexistent);

Update specified attributes on specified documents

Name of the index to be updated
Array of attribute name strings
A hash with key as document id, value as an array of new attribute values
If set, indicates that there is update of MVA attributes
If set, the update will silently ignore any warnings about trying to update a column which is not exists in current index schema.

Returns number of actually updated documents (0 or more) on success

Returns undef on failure

Usage example:

 $sph->UpdateAttributes("test1", [ qw/group_id/ ], { 1 => [ 456] }) );



Opens a persistent connection for subsequent queries.

To reduce the network connection overhead of making Sphinx queries, you can call $sph->Open(), then run any number of queries, and call $sph->Close() when finished.

Returns 1 on success, 0 on failure.



Closes a persistent connection.

Returns 1 on success, 0 on failure.


    $status = $sph->Status()
    $status = $sph->Status($session)

Queries searchd status, and returns a hash of status variable name and value pairs.

Returns undef on failure.



There is (or was) a bundled Sphinx.pm in the contrib area of the Sphinx source distribution, which was used as the starting point of Sphinx::Search. Maintenance of that version appears to have lapsed at sphinx-0.9.7, so many of the newer API calls are not available there. Sphinx::Search is mostly compatible with the old Sphinx.pm except:
On failure, Sphinx::Search returns undef rather than 0 or -1.
Sphinx::Search 'Set' functions are cascadable, e.g. you can do Sphinx::Search->new ->SetMatchMode(SPH_MATCH_ALL) ->SetSortMode(SPH_SORT_RELEVANCE) ->Query("search terms")

Sphinx::Search also provides documentation and unit tests, which were the main motivations for branching from the earlier work.


Jon Schutz



Please report any bugs or feature requests to "bug-sphinx-search at rt.cpan.org", or through the web interface at <http://rt.cpan.org/NoAuth/ReportBug.html?Queue=Sphinx-Search>. I will be notified, and then you'll automatically be notified of progress on your bug as I make changes.


You can find documentation for this module with the perldoc command.

    perldoc Sphinx::Search

You can also look for information at:


This module is based on Sphinx.pm (not deployed to CPAN) for Sphinx version 0.9.7-rc1, by Len Kranendonk, which was in turn based on the Sphinx PHP API.

Thanks to Alexey Kholodkov for contributing a significant patch for handling persistent connections.


Copyright 2015 Jon Schutz, all rights reserved.

This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License.