[DRILL-6976] SchemaChangeException happens when using split function in subquery if it returns empty result.

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.15.0
• Fix Version/s: 1.16.0
• Assignee: Bohdan Kazydub
• Reporter: Anton Gozhiy
• Created: 14/Jan/19 18:23
• Updated: 18/Mar/19 11:19
• Resolved: 12/Mar/19 17:52

Description

Query:

select substr(col, 2, 3) 
from (select split(n_comment, ' ') [3] col 
      from cp.`tpch/nation.parquet` 
      where n_nationkey = -1 
      group by n_comment 
      order by n_comment 
      limit 5);

Expected result:

+---------+
| EXPR$0  |
+---------+
+---------+

Actual result:

Error: SYSTEM ERROR: SchemaChangeException: Failure while trying to materialize incoming schema.  Errors:
 
Error in expression at index -1.  Error: Missing function implementation: [castVARCHAR(NULL-OPTIONAL, BIGINT-REQUIRED)].  Full expression: --UNKNOWN EXPRESSION--..

Fragment 0:0

Please, refer to logs for more information.

[Error Id: 86515d74-7b9c-4949-8ece-c9c17e00afc3 on userf87d-pc:31010]

  (org.apache.drill.exec.exception.SchemaChangeException) Failure while trying to materialize incoming schema.  Errors:
 
Error in expression at index -1.  Error: Missing function implementation: [castVARCHAR(NULL-OPTIONAL, BIGINT-REQUIRED)].  Full expression: --UNKNOWN EXPRESSION--..
    org.apache.drill.exec.physical.impl.project.ProjectRecordBatch.setupNewSchemaFromInput():498
    org.apache.drill.exec.physical.impl.project.ProjectRecordBatch.setupNewSchema():583
    org.apache.drill.exec.record.AbstractUnaryRecordBatch.innerNext():101
    org.apache.drill.exec.physical.impl.project.ProjectRecordBatch.innerNext():143
    org.apache.drill.exec.record.AbstractRecordBatch.next():186
    org.apache.drill.exec.physical.impl.BaseRootExec.next():104
    org.apache.drill.exec.physical.impl.ScreenCreator$ScreenRoot.innerNext():83
    org.apache.drill.exec.physical.impl.BaseRootExec.next():94
    org.apache.drill.exec.work.fragment.FragmentExecutor$1.run():297
    org.apache.drill.exec.work.fragment.FragmentExecutor$1.run():284
    java.security.AccessController.doPrivileged():-2
    javax.security.auth.Subject.doAs():422
    org.apache.hadoop.security.UserGroupInformation.doAs():1746
    org.apache.drill.exec.work.fragment.FragmentExecutor.run():284
    org.apache.drill.common.SelfCleaningRunnable.run():38
    java.util.concurrent.ThreadPoolExecutor.runWorker():1149
    java.util.concurrent.ThreadPoolExecutor$Worker.run():624
    java.lang.Thread.run():748 (state=,code=0)

Note: Filter "where n_nationkey = -1" doesn't return any rows. In case of " = 1", for example, the query will return result without error.

Workaround: Use cast on the split function, like

cast(split(n_comment, ' ') [3] as varchar)

Issue Links

• is related to: DRILL-6962 Function coalesce returns an Error when none of the columns in coalesce exist in a parquet file