You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@impala.apache.org by "Zoltán Borók-Nagy (Jira)" <ji...@apache.org> on 2023/11/17 14:30:00 UTC

[jira] [Created] (IMPALA-12568) Add conflictDetectionFilter for Iceberg DML statements

Zoltán Borók-Nagy created IMPALA-12568:
------------------------------------------

             Summary: Add conflictDetectionFilter for Iceberg DML statements
                 Key: IMPALA-12568
                 URL: https://issues.apache.org/jira/browse/IMPALA-12568
             Project: IMPALA
          Issue Type: Bug
            Reporter: Zoltán Borók-Nagy


Currently Impala is too rigorous when it checks for conflicting data or delete files.

Currently any concurrently added data/delete file conflict with the current DML operation, even if they touch different partitions.

Iceberg has [conflictDetectionFilter|https://iceberg.apache.org/javadoc/1.0.0/org/apache/iceberg/RowDelta.html#conflictDetectionFilter-org.apache.iceberg.expressions.Expression-]([Expression|https://iceberg.apache.org/javadoc/1.0.0/org/apache/iceberg/expressions/Expression.html] conflictDetectionFilter) API to relax this. I.e. we could only check for conflicting data/delete files that are in partitions that the current DML operation touches.

This way we could e.g. compact a partition while other partitions are being written.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)