Re: Group by query reports null pointer when unique key is not stored

[Christine Poerschke (BLOOMBERG/ LONDON)]

Hello.

I'd probably describe this as a bug.

w.r.t. documentation adjustments e.g. 
https://github.com/apache/solr/blob/main/solr/solr-ref-guide/modules/indexing-guide/pages/schema-elements.adoc
 and 
https://github.com/apache/solr/blob/main/solr/solr-ref-guide/modules/query-guide/pages/result-grouping.adoc
 is the source behind the Solr Ref Guide and changes can be contributed via 
pull requests, just like with code. For small edits perhaps just working in 
your browser might suit and/or for more comprehensive information please see 
the 
https://github.com/apache/solr/blob/main/CONTRIBUTING.md#contributing-documentation
 details.

Best wishes,
Christine

From: users@solr.apache.org At: 02/07/24 00:31:23 UTCTo:  users@solr.apache.org
Subject: Re: Group by query reports null pointer when unique key is not stored

I also got this exception before and in order to avoid reindexing TB's of data, 
had to resort to grouping via streaming expressions, which has its ups and 
downs. If it's technically infeasible to substitute docValues for this purpose 
(when useDocValuesAsStored:true), it would be nice if it was documented in 
"grouping" feature or advised in general schema design pages to make 
unique ID's always stored:true.

-ufuk yilmaz

From: rajani m 
Sent: Wednesday, February 7, 2024 3:18 AM
To: solr-user 
Subject: Group by query reports null pointer when unique key is not stored

Hi Solr Users,

  Group by query is failing with the following error message. It looks like
TopGroupsResultTransformer.java

retrieveDocument method is fetching doc based on id is calling a stored
field visitor method. I tried setting the "id" field to use doc values as
they are enabled "useDocValuesAsStored:true" but this didn't help. Any
alternative?

Would you call this a bug or improvement?

Error message -

java.lang.NullPointerException: Cannot invoke
"org.apache.lucene.index.IndexableField.stringValue()" because "f" is
null => java.lang.NullPointerException: Cannot invoke
"org.apache.lucene.index.IndexableField.stringValue()" because "f" is
null
at org.apache.solr.schema.FieldType.toExternal(FieldType.java:378)
java.lang.NullPointerException: Cannot invoke
"org.apache.lucene.index.IndexableField.stringValue()" because "f" is
null
at org.apache.solr.schema.FieldType.toExternal(FieldType.java:378) 
~[?:?]
at 
org.apache.solr.search.grouping.distributed.shardresultserializer.TopGroupsResul
tTransformer.serializeTopGroups(TopGroupsResultTransformer.java:238)
~[?:?]

Re: Group by query reports null pointer when unique key is not stored

[Robi Petersen]

Interesting, Thanks Christine. That's a great suggestion. There's so many
places where the reg guide could be improved...

Best
Robi

On Wed, Feb 7, 2024 at 2:20 AM Christine Poerschke (BLOOMBERG/ LONDON) <
cpoersc...@bloomberg.net> wrote:

> Hello.
>
> I'd probably describe this as a bug.
>
> w.r.t. documentation adjustments e.g.
> https://github.com/apache/solr/blob/main/solr/solr-ref-guide/modules/indexing-guide/pages/schema-elements.adoc
> and
> https://github.com/apache/solr/blob/main/solr/solr-ref-guide/modules/query-guide/pages/result-grouping.adoc
> is the source behind the Solr Ref Guide and changes can be contributed via
> pull requests, just like with code. For small edits perhaps just working in
> your browser might suit and/or for more comprehensive information please
> see the
> https://github.com/apache/solr/blob/main/CONTRIBUTING.md#contributing-documentation
> details.
>
> Best wishes,
> Christine
>
> From: users@solr.apache.org At: 02/07/24 00:31:23 UTCTo:
> users@solr.apache.org
> Subject: Re: Group by query reports null pointer when unique key is not
> stored
>
> I also got this exception before and in order to avoid reindexing TB's of
> data,
> had to resort to grouping via streaming expressions, which has its ups and
> downs. If it's technically infeasible to substitute docValues for this
> purpose
> (when useDocValuesAsStored:true), it would be nice if it was documented in
> "grouping" feature or advised in general schema design pages to make
> unique ID's always stored:true.
>
> -ufuk yilmaz
> 
> From: rajani m 
> Sent: Wednesday, February 7, 2024 3:18 AM
> To: solr-user 
> Subject: Group by query reports null pointer when unique key is not stored
>
> Hi Solr Users,
>
>   Group by query is failing with the following error message. It looks like
> TopGroupsResultTransformer.java
> <
> https://github.com/apache/solr/blob/main/solr/core/src/java/org/apache/solr/sea
>
> rch/grouping/distributed/shardresultserializer/TopGroupsResultTransformer.java#L
> 237C24-L237C40
> 
> >
> retrieveDocument method is fetching doc based on id is calling a stored
> field visitor method. I tried setting the "id" field to use doc values as
> they are enabled "useDocValuesAsStored:true" but this didn't help. Any
> alternative?
>
> Would you call this a bug or improvement?
>
> Error message -
>
> java.lang.NullPointerException: Cannot invoke
> "org.apache.lucene.index.IndexableField.stringValue()" because "f" is
> null => java.lang.NullPointerException: Cannot invoke
> "org.apache.lucene.index.IndexableField.stringValue()" because "f" is
> null
> at org.apache.solr.schema.FieldType.toExternal(FieldType.java:378)
> java.lang.NullPointerException: Cannot invoke
> "org.apache.lucene.index.IndexableField.stringValue()" because "f" is
> null
> at org.apache.solr.schema.FieldType.toExternal(FieldType.java:378)
> ~[?:?]
> at
>
> org.apache.solr.search.grouping.distributed.shardresultserializer.TopGroupsResul
> tTransformer.serializeTopGroups(TopGroupsResultTransformer.java:238)
> ~[?:?]
>
>
>