Great work Jorg!. We are going to go through the effort of upgrading to Lucene 4.0, and then restructure a bit the facet API code (to simplify writing custom ones for example). I would also say that, if its just for ordering, we should allow for a pluggable "sort" logic for the different terms facet, which ICU can add.
On Oct 19, 2012, at 9:49 PM, Jörg Prante firstname.lastname@example.org wrote:
Well, I had to rework the pull request - it wasn't working, but now it should. Sorry for the inconvenience.
Also I updated the download zip.
For users who prefer the non-ICU environment (JDK-based) for java.text.Collator based sorting on doc lists and facets, I issued pull requests:
On Wednesday, October 17, 2012 11:19:28 AM UTC+2, Tanguy wrote:
Sounds interesting for french as well, thanks Jorg
Le mercredi 17 octobre 2012 11:04:57 UTC+2, Jörg Prante a écrit :
just a quick note, I want to draw your attention to a little feature I just added to the ICU plugin. A new facet type "icu" allows sorting string entries in facets according to ICU collation rules.
For german users, the challenge is well-known in displaying names in facets that are sorted appropriately even if they have umlauts ("Telefonbuchsortierung").
Pull request here https://github.com/elasticsearch/elasticsearch-analysis-icu/pull/7
Preliminary ZIP as drop-in replacement for the ICU plugin