Thanks for bringing up performance perspective. Is there any bench mark on
join performance when number of shards is more than 10 where documents are
indexed based on router.field.
Are you suggesting instead of router.field go for streaming expressions or
use join with router.field and then go for
this will likely be "interesting" from a performance perspective. You
might try Streaming, especially StreamingExpressions and ParallelSQL
depending on what you need this for.
Best,
Erick
On Mon, May 22, 2017 at 12:05 AM, Damien Kamerman wrote:
> I use a router.field so docs that I join from/to
I use a router.field so docs that I join from/to are always in the same
shard. See
https://cwiki.apache.org/confluence/display/solr/Shards+and+Indexing+Data+in+SolrCloud#ShardsandIndexingDatainSolrCloud-DocumentRouting
There is an open ticket SOLR-8297
https://issues.apache.org/jira/browse/SOLR-8
Is there any possibility of supporting joins across multiple shards in near
future ? How to achieve the join when our data is spread-ed across multiple
shards. This is very much mandatory when we need to scale out.
Any workarounds if out-of-box possibility is not there ?
Thanks,
--
View th
Perfect !
Sorry I overlooked and missed "="
Thanks,
--
View this message in context:
http://lucene.472066.n3.nabble.com/Join-not-working-in-Solr-6-5-tp4336247p4336251.html
Sent from the Solr - User mailing list archive at Nabble.com.
Your join should be:
{!join from=id to=C_pid_s}
On 22 May 2017 at 14:07, mganeshs wrote:
> Hi,
>
> I have following records / documents with Parent entity
>
> id,type_s,P_hid_s,P_name_s,P_pid_s
> 11,PERSON,11,Parent1,11
>
> And following records / documents with child en
Hi,
I have following records / documents with Parent entity
id,type_s,P_hid_s,P_name_s,P_pid_s
11,PERSON,11,Parent1,11
And following records / documents with child entity
id,type_s,C_hid_s,C_name_s,C_pid_s
12,PERSON,12,Child2,11
13,PERSON,
