You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by "Alexey Vasyukov (JIRA)" <ji...@apache.org> on 2016/06/19 15:11:05 UTC

[jira] [Closed] (SOLR-9224) Nested objects for multilevel nesting

     [ https://issues.apache.org/jira/browse/SOLR-9224?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Alexey Vasyukov closed SOLR-9224.
---------------------------------
    Resolution: Not A Bug

My bad. Solved in the discussion at http://stackoverflow.com/questions/37905553/are-solr-nested-objects-broken-for-multilevel-nesting It's not obvious, but it works after all.

> Nested objects for multilevel nesting
> -------------------------------------
>
>                 Key: SOLR-9224
>                 URL: https://issues.apache.org/jira/browse/SOLR-9224
>             Project: Solr
>          Issue Type: Bug
>    Affects Versions: 5.5.1, 6.1
>            Reporter: Alexey Vasyukov
>
> I observe quite strange behaviour for multi-level nesting of SOLR objects using both SOLR 5.5.1 and SOLR 6.1.0 (SOLR Cloud to be specific, if it's important).
> I consider it to be a SOLR bug. However, my experience with nested objects is limited, so, I'd really appreciate any review from more experienced folks. The steps to reproduce the issue:
> 1. Start clean SOLR instance and create test collection:
> {code}
> ./bin/solr start -e cloud -noprompt
> ./bin/solr create_collection -c qa -shards 2 -p 8983
> {code}
> 2. Upload test data:
> {code}
>     <doc>
>         <field name="id">CONTAINER_1</field>
>         <field name="title">CONTAINER</field>
>         <doc>
>             <field name="id">CONTAINER_1_ITEM</field>
>             <field name="title">META</field>
>         </doc>
>         <doc>
>             <field name="id">CONTAINER_1_PERSON</field>
>             <field name="title">META</field>
>         </doc>
>         <doc>
>             <field name="id">DOC_1</field>
>             <field name="title">DOCUMENT</field>
>             <doc>
>                 <field name="id">DOC_1_ITEM</field>
>                 <field name="title">META</field>
>             </doc>
>             <doc>
>                 <field name="id">DOC_1_PERSON</field>
>                 <field name="title">META</field>
>             </doc>
>         </doc>
>         <doc>
>             <field name="id">DOC_2</field>
>             <field name="title">DOCUMENT</field>
>             <doc>
>                 <field name="id">DOC_2_ITEM</field>
>                 <field name="title">META</field>
>             </doc>
>             <doc>
>                 <field name="id">DOC_2_PERSON</field>
>                 <field name="title">META</field>
>             </doc>
>         </doc>
>     </doc>
> {code}
> 3. Query for:
> {code}
> {!child of=title:DOCUMENT}id:DOC_2
> {code}
> Everything works as expected, there are 2 docs in the response:
> {code}
>     [
>       {
>         "id":"DOC_2_ITEM",
>         "title":["META"]},
>       {
>         "id":"DOC_2_PERSON",
>         "title":["META"]}
>     ]
> {code}
> 4. Query for:
> {code}
> {!child of=title:DOCUMENT}id:DOC_1
> {code}
> See strange response with 4 docs instead of 2:
> {code}
>     [
>       {
>         "id":"CONTAINER_1_ITEM",
>         "title":["META"]},
>       {
>         "id":"CONTAINER_1_PERSON",
>         "title":["META"]},
>       {
>         "id":"DOC_1_ITEM",
>         "title":["META"]},
>       {
>         "id":"DOC_1_PERSON",
>         "title":["META"]}
>     ]
> {code}
> Any ideas on what is wrong? Any ideas regarding possible workaround?
> I'm absolutely positive about debugging-fixing-contributing myself, but I really need initial entry points for digging into SOLR sources.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

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