[jira] [Resolved] (SOLR-10221) Search for "jac*" or "jacob*" does not work for "jacobus" when it is part of long lists of synonyms

["Erick Erickson (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/SOLR-10221?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Erick Erickson resolved SOLR-10221.
-----------------------------------
    Resolution: Invalid

First of all, please raise questions like this on the user's list first before raising a JIRA to see if the behavior you see is really a bug or not.

In this case it's not. The problem you're having is that wildcards do not go through synonym expansion because synonym filters are not "MultiTermAware". That is, they may produce more than one token on output per input token, so there is no "correct" behavior. See: https://lucidworks.com/2011/11/29/whats-with-lowercasing-wildcard-multiterm-queries-in-solr/

> Search for "jac*" or "jacob*" does not work for "jacobus" when it is part of long lists of synonyms
> ---------------------------------------------------------------------------------------------------
>
>                 Key: SOLR-10221
>                 URL: https://issues.apache.org/jira/browse/SOLR-10221
>             Project: Solr
>          Issue Type: Bug
>      Security Level: Public(Default Security Level. Issues are Public) 
>    Affects Versions: 4.10.4, 6.2
>            Reporter: Alexandar Mitsev
>
> We have a field with SynonymFilterFactory for dutch synonyms. Both on index and on query. The word "jacob" and other spellings of "jacob" are presented in 3 very long list of synonyms.
> We have the word "jacobus" indexed in there.
> Search for "jac*", "jacob*" and etc. does not work. It does work for other words. And it does work if the synonyms are not used. Or if the synonyms are only used on query time.
> The two lists of synonyms which somehow brake it are:
> {noformat}
> cobes, cobis, cobus, coobes, iakobus, iakop, ijacob, ijacobis, ijacobus, ijapick, jaacke, jaacob, jaaipik, jaakes, jaakob, jaakoob, jaap, jaapik, jabec, jac., jaccob, jaces, jachop, jacke, jackes, jackob, jackop, jacob, jacobes, jacobis, jacobp, jacobs, jacobus, jacoch, jacoob, jacop, jacq, jacque, jacques, jacquis, jacub, jacus, jaecke, jaeckob, jaecob, jaecques, jaeke, jaekes, jaekob, jaep, jak, jak., jake, jakis, jakke, jakkob, jakkop, jakob, jakobes, jakobi, jakobis, jakobje, jakobus, jakoob, jakoobes, jakop, jakques, jakus, james, japek, japick, japijck, japijk, japik, japje, jappe, jappik, japyck, japyk, jaques, jaquez, jaquis, jeems, jeppik, kobbis, kobes, kobis, kobise, kobus, kobuse, koobes, koobis, koobus, koos, yacob, yacobis, yacobus, yapick => JAKOB
> cobes, cobis, cobus, coobes, coobus, iakobus, iakops, ijacobs, ijapicks, jaabse, jaackis, jaacobs, jaakes, jaakobs, jaakoobs, jaapiks, jaaps, jabex, jabics, jabiks, jac., jacacobs, jaccobs, jaccobsdr, jaccobsz, jaces, jachops, jacis, jackes, jackobes, jackobhs, jackobs, jackops, jacob, jacobdr, jacobes, jacobesz, jacobi, jacobij, jacobis, jacobo, jacobpsz, jacobs, jacobsd, jacobsdr, jacobse, jacobsen, jacobsens, jacobss, jacobsz, jacobszen, jacobszn, jacobszoon, jacobus, jacobussen, jacobusz, jacoby, jacobz, jacobzn, jacochs, jacoobes, jacoobs, jacopdr, jacops, jacopsdr, jacopsz, jacopszn, jacos, jacques, jac.s, jacubs, jacus, jaeckes, jaeckesdr, jaeckobs, jaeckops, jaecobs, jaecobsz, jaekes, jaekobs, jaeques, jakes, jakis, jakkobs, jakkops, jakob, jakobes, jakobessen, jakobij, jakobis, jakobjes, jakobs, jakobsdr, jakobse, jakobsen, jakobus, jakoby, jakobz, jakoobs, jakoops, jakop, jakops, jakques, jaks, jakus, james, jaobs, japeks, japicks, japicksdr, japics, japicx, japiks, japikx, japix, jappedr, jappes, jappesdr, jappesz, jappeszn, jappezn, jappiks, jappis, jaques, jaquesdr, jaquis, jaquisdr, jaquusdr, jcobs, jeppicx, jeppiks, kobbis, kobes, kobijs, kobis, kobises, kobus, kobuses, kobys, koobes, koobis, koobus, yacobs, yapicks => JAKOBS
> {noformat}



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org