http://wiki.apache.org/solr/AnalyzersTokenizersTokenFilters#solr.WordDel...
this may be useful in some cases.
| Comment | File | Size | Author |
|---|---|---|---|
| #7 | 824154-preserveOriginal-7.patch | 2.43 KB | pwolanin |
| #6 | 824154-preserveOriginal-6.patch | 2.42 KB | pwolanin |
| #2 | 824154-preserveOriginal-2.patch | 2.42 KB | pwolanin |
| #1 | 824154-preserveOriginal-1.patch | 2 KB | pwolanin |
Comments
Comment #1
pwolanin commentedadded the 1 attribute, plus reformated the rest.
Comment #2
pwolanin commentedshould bump the version in the schema name too
Comment #3
robertdouglass commentedWe're considering making this change in both branches. Anyone have opinions? It's a schema change on a 1.0 release...
Comment #4
pwolanin commentedThe motivation:
1) this probably improves search accuracy e.g. for hyphenated words, camel case words, etc.
2) If you use the standard query parser or the edismax parser, wildcard queries are not analyzed (according to Erik Hatcher). Thus, a wildcard search for a hyphenated word or word otherwise split by the word delimited filter will not match.
Comment #5
jpmckinney commentedSounds good to me.
Comment #6
pwolanin commentedoops we already put the schema to drupal-0.9.5, so this should go to 0.9.6 - or 1.0 ?
-Peter
Comment #7
pwolanin commenteddiscovered that this schmea has diverged as of: #382878: Render $node->teaser before sending it to Solr
Adding 1 to the version name for 6.x-2.x, HEAD.
Comment #8
pwolanin commentedcommitted #6 to 5.x-2.x, and 6.x-1.x, #7 to 6.x-2.x and HEAD (7.x)