I believe I may have solved this. After a more careful reading of
SOLR-1653, I noticed that they referred to another filter. I changed my
configuration from /solr/.PatternReplaceCharFilterFactory to
/solr/.PatternReplaceFilterFactory and updated the XML syntax
appropriately, and it looks OK now. This filter is not mentioned on the
wiki page dealing with analyzers, which is why I did not use it from the
start. When I searched that page for regex, the CharFilter was the only
one that came up.
On 8/31/2010 8:29 AM, Shawn Heisey wrote:
I didn't give any particulars about my setup, sorry about that. This
is branch_3x rev 990625, downloaded two days ago. It passed all unit
tests.
Linux idxst9-b 2.6.32-bpo.5-amd64 #1 SMP Fri Jun 11 08:42:31 UTC 2010
x86_64 GNU/Linux
Java(TM) SE Runtime Environment (build 1.6.0_20-b02)
Java HotSpot(TM) 64-Bit Server VM (build 16.3-b01, mixed mode)
On 8/31/2010 8:23 AM, Shawn Heisey wrote:
I am trying to use PatternReplaceCharFilterFactory (SOLR-1653) to
strip leading and trailing punctuation from terms. It's not
working. This was previously discussed here as part of something I
was trying with WordDelimiterFilterFactory, but I think it needs its
own thread now.