[SOLR-7669] Add SelectStream to Streaming API

• Type: Improvement
• Status: Closed
• Priority: Minor
• Resolution: Implemented
• Fix Version/s: 6.0
• Component/s: SolrJ
• Labels: Streaming
• Assignee: Dennis Gove
• Reporter: Dennis Gove
• Created: 12/Jun/15 02:06
• Updated: 09/May/16 18:44
• Resolved: 12/Nov/15 04:39

Description

Adds a new stream called SelectStream which can be used for two purpose.
1. Limit the set of fields included in an outgoing tuple to remove unwanted fields
2. Provide aliases for fields. With this it acts as an alternative to the CloudSolrStream's 'aliases' option.

For example, in a simple case

select(
  id, 
  fieldA_i as fieldA, 
  fieldB_s as fieldB,
  search(collection1, q="*:*", fl="id,fieldA_i,fieldB_s", sort="fieldA_i asc, fieldB_s asc, id asc")
)

This can also be used as part of complex expressions to help keep track of what is being worked on. This is particularly useful when merging/joining multiple collections which share field names. For example, the following results in a set of tuples including only the fields id, left.ident, and right.ident even though the total set of fields required to perform the search and join is much larger than just those three fields.

select(
  id, left.ident, right.ident,
  innerJoin(
    select(
      id, join1_i as left.join1, join2_s as left.join2, ident_s as left.ident,
      search(collection1, q="side_s:left", fl="id,join1_i,join2_s,ident_s", sort="join1_i asc, join2_s asc, id asc")
    ),
    select(
      join3_i as right.join1, join2_s as right.join2, ident_s as right.ident,
      search(collection1, q="side_s:right", fl="join3_i,join2_s,ident_s", sort="join3_i asc, join2_s asc"),
    ),
    on="left.join1=right.join1, left.join2=right.join2"
  )
)

This depends on SOLR-7584.

Issue Links

• is related to: SOLR-8125 Umbrella ticket for Streaming and SQL issues