#016127: Searching with numeric values

Both issues are caused by "bugs" in Zend_Search_Lucene.

For ::INT: If a part of a query does not match anything, it is considered irrelevant, and that may cause the whole query to be reduced into a null-query, which then returns all results.

For ::STRING: The query tokenizer in ZSL just ignores numeric values. A patch for this is simple, but still should be fixed by the Zend-guys. I will open bug reports for both upstream, and link those issues here. The simple patch for ::STRING is:

Index: /Zend/Search/Lucene/Analysis/Analyzer/Common/Text.php
===================================================================
--- Zend/Search/Lucene/Analysis/Analyzer/Common/Text.php    (revision 19211)
+++ Zend/Search/Lucene/Analysis/Analyzer/Common/Text.php    (working copy)
@@ -75,7 +75,7 @@
 
 
         do {
-            if (! preg_match('/[a-zA-Z]+/', $this->_input, $match, PREG_OFFSET_CAPTURE, $this->_position)) {
+            if (! preg_match('/[a-zA-Z0-9]+/', $this->_input, $match, PREG_OFFSET_CAPTURE, $this->_position)) {
                 // It covers both cases a) there are no matches (preg_match(...) === 0)
                 // b) error occured (preg_match(...) === FALSE)
                 return null;

#264636 by Kore Nordmann on February 12th, 2010 [Permanent Link]

Issue http://framework.zend.com/issues/browse/ZF-5236 showed a workaround, so it has been fixed and tested in revision #11371.

#264638 by Kore Nordmann on February 12th, 2010 [Permanent Link]

Thanks for the fix.

On a somewhat related note, wouldn't it be better to use the case insensitive version instead?

new Zend_Search_Lucene_Analysis_Analyzer_Common_Utf8Num_CaseInsensitive()

instead of

new Zend_Search_Lucene_Analysis_Analyzer_Common_Utf8Num()

#264682 by Nathan Guse on February 15th, 2010 [Permanent Link]

Reopening as it it seems nobody has noticed my reply about using Zend_Search_Lucene_Analysis_Analyzer_Common_Utf8Num_CaseInsensitive()

#264881 by Nathan Guse on February 25th, 2010 [Permanent Link]

Switched to that, even it will only affect ASCII chars, and no other Unicode-characters. Still, it won't hurt, I guess.

#265110 by Kore Nordmann on March 5th, 2010 [Permanent Link]

See above.

#265111 by Kore Nordmann on March 5th, 2010 [Permanent Link]