[jira] [Created] (DRILL-5021) ExternalSortBatch redundantly redefines the batch schema

["Paul Rogers (JIRA)" <ji...@apache.org>]

Paul Rogers created DRILL-5021:
----------------------------------

             Summary: ExternalSortBatch redundantly redefines the batch schema
                 Key: DRILL-5021
                 URL: https://issues.apache.org/jira/browse/DRILL-5021
             Project: Apache Drill
          Issue Type: Bug
    Affects Versions: 1.8.0
            Reporter: Paul Rogers
            Priority: Minor


Much code in the {{ExternalSortBatch}} (ESB) deals with building vector batches and schemas. However, ESB cannot handle schema changes. The only valid schema difference is the same field path in a different position in the vector array. Given this restriction, the code can be simplified (and sped up) by exploiting the fact that all batches are required to have the same conceptual schema (same set of fields, but perhaps in different vector order) and most probably, the same physical schema (same fields and same vector order.) Note that, because of the way that the {{getValueVectorId()}} method works, each lookup of a value vector is an O(n) operation, so that each remapping of vectors is O(n^2).




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