Also note that Solr field names really shouldn't contain the period.
>From the ref guide:
"Field names should consist of alphanumeric or
underscore characters only and not start with a digit..."
names with dots won't be rejected, but aren't guaranteed to
be supported in all situations either.
Best,
Erick
On Tue, Jan 16, 2018 at 3:21 AM, Mikhail Khludnev <m...@apache.org> wrote:
> I notice the space after the 2nd +. It's not a really big issue, but it
> should be recognized as a syntax error (I believe), it might mean that your
> http layer may grab pluses and turn them into spaces, try to replace them
> to %2b.
> Regarding the exception, this is what I tried to protect from with the
> sibling +type:address, and you can literally follow the advice from
> exception message, however it just proves hypothesis above.
>
> On Tue, Jan 16, 2018 at 12:11 PM, Dariusz Wojtas <dwoj...@gmail.com> wrote:
>
>> Please see the answers below.
>>
>> On Tue, Jan 16, 2018 at 8:40 AM, Mikhail Khludnev <m...@apache.org> wrote:
>>
>> > On Tue, Jan 16, 2018 at 10:30 AM, Dariusz Wojtas <dwoj...@gmail.com>
>> > wrote:
>> >
>> > > Hello Mikhail,
>> > >
>> > > I've tried it and the query executes, but it does not treat strdist()
>> as
>> > a
>> > > function to execute.
>> > > Looks like each part of the function - it's name and parameters - are
>> > > treated as keywords to search for against the default field.
>> > >
>> > Can you post the exact observations, rather than interpretation?
>> >
>>
>> Yes,
>> Here is the output of the original query proposed, where I believe strdist
>> call is not treated as a function. keyword1 if the default search field:
>>
>> <str name="parsedquery">
>> firstName:Adam AllParentsAware(ToParentBlockJoinQuery (type:address
>> keyword1:strdic (keyword1:shakespear keyword1:address.streeo
>> keyword1:edit)))
>> </str>
>> This is not a function interpretation call.
>>
>>
>> > If I try something different:
>> > >
>> > > q=+firstName:Adam +{!parent which=type:record
>> > v=$chq}&chq=+type:address
>> > > +{!func}strdist('Shakespeare',address.street, edit)
>> > >
>> > > then I get exception:
>> > > org.apache.solr.search.SyntaxError: Missing end to unquoted value
>> > > starting at 37 str='strdist('Shakespeare',address.street,'
>> > >
>> > This particular query failed because of the space. Here is my pet peeve
>> in
>> > Solr: the syntax {!foo} captures whole string if it's in beginning of the
>> > string, but in the middle of the string it captures only substring until
>> > the first space.
>> > So, after removing space it should work fine. Another potential issues
>> are:
>> > single quotes (do they it ever supported?), and the dot in the fieldname
>> > (you never know).
>>
>>
>>
>> Yes, it's been about the space first, then about the dot in the name of the
>> nested field 'address.street'.
>> After many failed attempts with various escaping modes, I have executed it
>> with double parameter dereferencing, but have faced another issue:
>>
>> q=+firstName:Adam + {!parent which='type:record' score='max' v=$chq1}
>> chq1=+type:address +{!func v='$chq2'}
>> chq2=strdist('Shakespeare',address.street,edit)
>>
>> java.lang.IllegalStateException: Child query must not match same docs with
>> parent filter. Combine them as must clauses (+) to find a problem doc.
>> docId=5, class org.apache.lucene.search.DisjunctionSumScorer at
>> org.apache.lucene.search.join.ToParentBlockJoinQuery$BlockJoinScorer.
>> setScoreAndFreq(ToParentBlockJoinQuery.java:327)
>> at
>> org.apache.lucene.search.join.ToParentBlockJoinQuery$
>> BlockJoinScorer.score(ToParentBlockJoinQuery.java:286)
>> at
>> org.apache.lucene.search.BooleanScorer$OrCollector.
>> collect(BooleanScorer.java:142)
>> at
>> org.apache.lucene.search.Weight$DefaultBulkScorer.
>> scoreRange(Weight.java:208)
>> at org.apache.lucene.search.Weight$DefaultBulkScorer.
>> score(Weight.java:195)
>> at
>> org.apache.lucene.search.BooleanScorer$BulkScorerAndDoc.score(
>> BooleanScorer.java:61)
>> at
>> org.apache.lucene.search.BooleanScorer.scoreWindowIntoBitSetAndReplay
>> (BooleanScorer.java:213)
>> at
>> org.apache.lucene.search.BooleanScorer.scoreWindowMultipleScorers(
>> BooleanScorer.java:260)
>> at
>> org.apache.lucene.search.BooleanScorer.scoreWindow(BooleanScorer.java:305)
>> at org.apache.lucene.search.BooleanScorer.score(BooleanScorer.java:317) at
>> org.apache.lucene.search.BulkScorer.score(BulkScorer.java:39) at
>> org.apache.lucene.search.IndexSearcher.search(IndexSearcher.java:658) at
>> org.apache.lucene.search.IndexSearcher.search(IndexSearcher.java:462) at
>> org.apache.solr.search.SolrIndexSearcher.buildAndRunCollectorChain(
>> SolrIndexSearcher.java:215)
>> at
>> org.apache.solr.search.SolrIndexSearcher.getDocListNC(
>> SolrIndexSearcher.java:1591)
>> at
>> org.apache.solr.search.SolrIndexSearcher.getDocListC(
>> SolrIndexSearcher.java:1408)
>> at
>> org.apache.solr.search.SolrIndexSearcher.search(
>> SolrIndexSearcher.java:575)
>> at
>>
>>
>> Any hint now?
>> Mo goal is to return the result of the strdist() function on a nested
>> document field.
>>
>> Best regards,
>> Dariusz Wojtas
>>
>>
>> >
>> > > Best regards,
>> > > Dariusz Wojtas
>> > >
>> > >
>> > >
>> > >
>> > > On Tue, Jan 16, 2018 at 4:04 AM, Mikhail Khludnev <m...@apache.org>
>> > wrote:
>> > >
>> > > > Hello, Dariusz.
>> > > >
>> > > > It should be something like
>> > > > q=+firstName:Adam +{!parent which=type:record
>> > > > v=$chq}&chq=+type:address +strdist('Shakespeare',
>> > > > address.street, edit)
>> > > > post exception if it doesn't work.
>> > > >
>> > > > On Tue, Jan 16, 2018 at 1:39 AM, Dariusz Wojtas <dwoj...@gmail.com>
>> > > wrote:
>> > > >
>> > > > > Hi,
>> > > > >
>> > > > > Is it possible to use the strdist() function to return distance on
>> > the
>> > > > > child document field?
>> > > > > Let's say I have:
>> > > > >
>> > > > > <doc>
>> > > > > <str name="id">1</str>
>> > > > > <str name="type">record</str>
>> > > > > <str name="firstName">Adam</str>
>> > > > > <doc>
>> > > > > <str name="id">A1</str>
>> > > > > <str name="type">address</str>
>> > > > > <str name="address.type">business</str>
>> > > > > <str name="address.street">Shakespeare</str>
>> > > > > </doc>
>> > > > > <doc>
>> > > > > <str name="id">A2</str>
>> > > > > <str name="type">address</str>
>> > > > > <str name="address.type">correspondence</str>
>> > > > > <str name="address.street">Baker Street</str>
>> > > > > </doc>
>> > > > > </doc>
>> > > > >
>> > > > > What I want to do is to search for documents:
>> > > > > type:record
>> > > > > firstName:Adam
>> > > > > and return max strdist('Shakespeare', address.street, edit) as
>> the
>> > > > > resulting score?
>> > > > >
>> > > > > or
>> > > > > type:record
>> > > > > firstName:Adam
>> > > > > and return max strdist('Shakespeare', address.street, edit) of
>> > > > > "address.type:business" as the resulting score?
>> > > > >
>> > > > > I am trying with the {!parent} mode and {!function}, various
>> > > > combinations.
>> > > > > But I do not get what I'd expect.
>> > > > >
>> > > > > Best regards,
>> > > > > Dariusz Wojtas
>> > > > >
>> > > >
>> > > >
>> > > >
>> > > > --
>> > > > Sincerely yours
>> > > > Mikhail Khludnev
>> > > >
>> > >
>> >
>> >
>> >
>> > --
>> > Sincerely yours
>> > Mikhail Khludnev
>> >
>>
>
>
>
> --
> Sincerely yours
> Mikhail Khludnev
