Support for Graph parser with multiple shards

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Support for Graph parser with multiple shards

sambasivarao giddaluri
Hi all ,
We have documents to type which parent , child, grandchild and each child
document has a reference field to parent doc and grandchild document has
reference fields to child doc and parent doc .

and each document has multiple fields

ex: few fields on parent doc are age , gender , name and similarly on child
and grandchild doc we have different fields

few of our use cases are

1) Get all parents documents where child documents matches condition (ex:
child.age>20) . this should give all the parent documents with all fields
2) Get all parents documents where parent condition matches,child condition
matches and grand child condition matches
3) Get all child documents where parent condition matches,child condition
matches and grand child condition matches


we also had few other complex use case with pagination and *all were
achievable by graph parser .*

1) But we are stuck as graph parser is not applicable with multi shard in
cloud mode
2) gatherNodes does not work with pagination and also it gathers only a
single field and we had to run another query to get all the fields on the
doc and  as it is streaming we had to pass sort field which does not work
if we are looking for relevancy.

Any plans to implement graph parser with multiple shards


I did went through patch  this
https://issues.apache.org/jira/browse/SOLR-8176 but not working as it is
getting exceptions with kafka.

Regards
sam