(PECL sphinx >= 0.1.0)
SphinxClient::query — Execute search query
Connects to searchd server, runs the given search query with the current settings, obtains and returns the result set.
An index name (or names).
On success, SphinxClient::query() returns a list of found matches and additional per-query statistics. The result set is a hash utilize other structures instead of hash) with the following keys and values:
|"matches"||An array with found document IDs as keys and their weight and attributes values as values|
|"total"||Total number of matches found and retrieved (depends on your settings)|
|"total_found"||Total number of found documents matching the query|
|"words"||An array with words (case-folded and stemmed) as keys and per-word statistics as values|
|"error"||Query error message reported by searchd|
|"warning"||Query warning reported by searchd|
There is also a "status" key showing in what status query has ended.
Here are codes, taken from searchd.cpp:
/// known status return codes
SEARCHD_OK = 0, ///< general success, command-specific reply follows
SEARCHD_ERROR = 1, ///< general failure, error message follows
SEARCHD_RETRY = 2, ///< temporary failure, error message follows, client should retry later
SEARCHD_WARNING = 3 ///< general success, warning message and command-specific reply follow
It appears that Query() (or the Sphinx program in general) does have some kind of limit imposed on the maximum length of a query submission. Stumbled across this limitation when trying to perform some searches based on content from user-submitted e-mails.
In looking at the returned array, the 'words' portion only accounts for the first 8 or 9 words of the long query. I'm assuming that the rest of the string is being truncated. The number of accepted words appears to change based on the total length of the words. Just a heads-up.