Re: [CLucene-dev] MultiFieldQueryParser - compared to Java

[Manuel F. <M.F...@ca...>]

A send the mail to early :)
The queries are incorrect in the mail:

Here are the correct :

>From Java: +(CNS_CAT:testcat CNS_DAT:testcat) +(CNS_CAT:barfoo CNS_DAT:barfoo)

>From C:  (+CNS_CAT:testcat +CNS_DAT:testcat) (+CNS_CAT:barfoo +CNS_DAT:barfoo)


Sorry
Manuel.

Von: Manuel Freiholz [mailto:M.F...@ca...]
Gesendet: Montag, 22. Juni 2009 10:17
An: clu...@li...
Betreff: [CLucene-dev] MultiFieldQueryParser - compared to Java

Hi,

i have an index with the following fields and data (clucene 0.9.21b):

| CNS_PATHID | CNS_CAT   | CNS_DAT
-------------------------------------------
| 1          | testcat   | foobar
| 2          | testcat   | barfoo
| 3          | testcat2  | boingflip

(Default operator AND)

When I search for "testcat barfoo" over the fields CNS_CAT and CNS_DAT
I want the result "2". I use the MultiFieldQueryParser to generate a Query object from the user query.

If I do this in Java, everything works.
Generated query from Java: +(CNS_CAT:testcat CNS_CAT:testcat) +(CNS_DAT: barfoo CNS_DAT:barfoo)
The query means, that the value "testcat" and "barfoo" MUST exist in ANY of the given fields.

If I do this with C, I get the following generated query:
Generated query from C: (+CNS_CAT:testcat +CNS_DAT:testcat) (
The query means, that the value "testcat" and "barfoo" MUST exist in ALL of the given fields.

It also seems like, that the change with "setDefaultOperator" has no effect.
I hardcoded it into the QueryParserBase constructor to have an effect.

Where is my mistake?^^

Thank you very much.
Manuel.