[IMPALA-11529] FILE__POSITION virtual column for ORC

• Type: New Feature
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Component/s: Backend
• Labels: iceberg
• Epic Color: ghx-label-12
• Assignee: Gabor Kaszab
• Reporter: Gabor Kaszab
• Created: 24/Aug/22 13:23
• Updated: 07/Sep/23 12:03
• Resolved: 01/Sep/22 09:13

Description

https://gerrit.cloudera.org/#/c/18704/ implemented the support of FILE_POSITION virtual column for Parquet files. This ticket covers the same but for ORC files.

Note, ORC scanner already has a functionality to calculate the position of a row in a file, in an optimal scenario it only has to be re-used for the FILE_POSITION virtual column.
see: https://github.com/apache/impala/blob/8e0482294975352d3d34d75adb50602d85b3c155/be/src/exec/orc/hdfs-orc-scanner.h#L275

Issue Links

• is duplicated by: IMPALA-11488 Add virtual column FILE__POSITION for ORC tables