bq: To me this seems like a design flaw. The Solr fieldtypes seem like they
allow a developer to create types that should handle wildcards
intelligently.
Well, that's pretty impossible. WordDelimiter(Graph)FilterFactory is a
case in point. It's designed to break up on
uppercase/lowercase/numeric/n
It doesn't seem to matter what you do in the query analyzer, if you have a
wildcard, it won't use it. Which is exactly the behavior I observed.
the solution was to set preserveOriginal="1" and change the etl process to
not strip the dashes, letting the index analyzer do that. We have a lot of
lega
Webster, did you try escaping the special character (assuming you did not
do what Shawn did by replacing - with some other text and your indexed
tokens have -)?
On Thu, Jul 27, 2017 at 12:03 PM, Webster Homer
wrote:
> Shawn,
> Thank you for that. I didn't know about that feature of the WDF. It d
Shawn,
Thank you for that. I didn't know about that feature of the WDF. It doesn't
help my situation but it's great to know about.
Googling solr wildcard searches I found this link
http://lucene.472066.n3.nabble.com/Wildcard-search-not-working-with-search-term-having-special-characters-and-digits-t
The Admin/Analysis page is useful here. It'll show you what each bit
of your query analysis chain does and may well point you to the part
of the chain that's the problem.
Best,
Erick
On Wed, Jul 26, 2017 at 11:33 AM, Webster Homer wrote:
> checked the Pattern Replace it's OK. Can't use the prese
checked the Pattern Replace it's OK. Can't use the preserve original since
it preserves the hyphens too, which I don't want. It would be best if it
didn't touch the * at all
On Wed, Jul 26, 2017 at 1:30 PM, Saurabh Sethi
wrote:
> My guess is PatternReplaceFilterFactory is most likely the cause.
My guess is PatternReplaceFilterFactory is most likely the cause.
Also, based on your query, you might want to set preserveOriginal=1
You can take one filter out at a time and see which one is altering the
query.
On Wed, Jul 26, 2017 at 11:13 AM, Webster Homer
wrote:
> 1. KeywordTokenizer - we
1. KeywordTokenizer - we want to treat the entire field as a single term to
parse
2. preserveOriginal = "0" Thought about changing this to 1
3. 6.2.2
This is the fieldtype
1. What tokenizer are you using?
2. Do you have preserveOriginal="1" flag set in your filter?
3. Which version of solr are you using?
On Wed, Jul 26, 2017 at 10:48 AM, Webster Homer
wrote:
> I have several fieldtypes that use the WordDelimiterFilterFactory
>
> We have a fieldtype for cas numbers
I have several fieldtypes that use the WordDelimiterFilterFactory
We have a fieldtype for cas numbers. which look like 1234-12-1, numbers
separated by hyphens, users often leave out the hyphens and either use
spaces or just string the numbers together.
The WDF seemed like a great solution especia
10 matches
Mail list logo
