You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hive.apache.org by "Krisztian Kasa (Jira)" <ji...@apache.org> on 2021/09/06 06:26:00 UTC

[jira] [Resolved] (HIVE-25406) Fetch writeId from insert-only transactional tables

     [ https://issues.apache.org/jira/browse/HIVE-25406?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Krisztian Kasa resolved HIVE-25406.
-----------------------------------
    Resolution: Fixed

Pushed to master. Thanks [~kgyrtkirk] for review.

> Fetch writeId from insert-only transactional tables
> ---------------------------------------------------
>
>                 Key: HIVE-25406
>                 URL: https://issues.apache.org/jira/browse/HIVE-25406
>             Project: Hive
>          Issue Type: Improvement
>          Components: ORC, Parquet, Reader, Vectorization
>            Reporter: Krisztian Kasa
>            Assignee: Krisztian Kasa
>            Priority: Major
>              Labels: pull-request-available
>          Time Spent: 20m
>  Remaining Estimate: 0h
>
> When generating plan for incremental materialized view rebuild a filter operator is inserted on top of each source table scans. The predicates contain a filter for writeId since we want to get all the rows inserted/deleted from the source tables since the last rebuild only.
> WriteId is part of the ROW_ID virtual column and only available for fully-ACID ORC tables.
> The goal of this jira is to populate a writeId when fetching from insert-only transactional tables.
> {code:java}
> create table t1(a int, b int) clustered by (a) into 2 buckets stored as orc TBLPROPERTIES ('transactional'='true', 'transactional_properties'='insert_only');
> ...
> SELECT t1.ROW__ID.writeId, a, b FROM t1;
> {code}



--
This message was sent by Atlassian Jira
(v8.3.4#803005)