[hibernate-issues] [Hibernate-JIRA] Commented: (HSEARCH-960) Index.UN_TOKENIZED overrides other tokenized fields that share the same name

[Sanne Grinovero (JIRA)]

    [ http://opensource.atlassian.com/projects/hibernate/browse/HSEARCH-960?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=44175#comment-44175 ] 

Sanne Grinovero commented on HSEARCH-960:
-----------------------------------------

Hi, thank you for the explanation.
Still you might have a problem: let's say you have an entity with the two fields annotated as in your description, having the values "ID5" and "Title of this book", and you use an Analyzer which includes lowercasing and removal of digits. Now during the query, you'll have to apply the same analyzer - which one? If you process the user input with the same analyzer your users will never be able to find "ID5", while if you don't process them they will need to type exactly "Title of this book", matching the same case, same whitespace and all terms in exactly the same order.

So in both cases you'll have elements in the index which can't be found and can't possibly produce hits - you would be better off with not indexing them at all.

Indexing multiple properties in the same field is indeed useful for the reasons as you say, but I'm afraid that to be useful they should all use the same Analyzer and Index options.

I think we should throw an exception or log a warning - unless someone can make a good example of a valid combination of a mismatched set of options.

Also.. why do you have users search for ID fields?

> Index.UN_TOKENIZED overrides other tokenized fields that share the same name
> ----------------------------------------------------------------------------
>
>                 Key: HSEARCH-960
>                 URL: http://opensource.atlassian.com/projects/hibernate/browse/HSEARCH-960
>             Project: Hibernate Search
>          Issue Type: Bug
>          Components: mapping
>    Affects Versions: 3.4.0.Final
>         Environment: 3.4.0 Final
>            Reporter: John-Michael Au
>              Labels: annotations, bug, override, tokenized, un_tokenized
>             Fix For: 3.4.2, 4.0.0.CR2
>
>
> Marking one field as un-tokenized causes all other fields with the same names to be un-tokenized.
> i.e.
> {code}
> @Field(name = "simple_search", index = Index.UN_TOKENIZED, store = Store.NO)
> private String string;
> @Field(name = "simple_search", index = Index.TOKENIZED, store = Store.NO)
> private String string2;
> {code}
> The resulting behaviour is that "simple_search" will be made up of un-tokenized 'string' and 'string2' values, even though 'string2' was specified to be tokenized.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira