hive-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alan Gates (JIRA)" <>
Subject [jira] [Commented] (HIVE-18124) clean up isAcidTable() API vs isInsertOnlyTable()
Date Mon, 18 Dec 2017 21:05:00 GMT


Alan Gates commented on HIVE-18124:

Checking that I have this straight:  acid = original acid, support update/delete/merge, isInertOnly
= micromanaged (aka 1/4 acid), transactional = acid || isInsertOnly.  Is that right?

Assuming I have that right, the changes seem fine.  We should just write that down somewhere,
preferably in the AcidUtils file, so future maintainers can keep up with what's going on.
 If it's already written somewhere else just a pointer to it in AcidUtils is fine.

With that caveat, +1.

>  clean up isAcidTable() API vs isInsertOnlyTable()
> --------------------------------------------------
>                 Key: HIVE-18124
>                 URL:
>             Project: Hive
>          Issue Type: Bug
>          Components: Transactions
>    Affects Versions: 3.0.0
>            Reporter: Eugene Koifman
>            Assignee: Eugene Koifman
>         Attachments: HIVE-18124.01.patch, HIVE-18124.02.patch, HIVE-18124.03.patch, HIVE-18124.04.patch,
> With the addition of MM tables (_AcidUtils.isInsertOnlyTable(table)_) the methods in
AcidUtils and dependent places are very muddled.  There are now a number of places where we
have something like _isAcidTable = AcidUtils.isFullAcidTable(table)_  and a later getter
> _boolean isAcidTable() \{ return isAcidTable;\}_
> Need to clean it up so that there is a isTransactional(Table) that checks transactional=true
setting and isAcid(Table) to mean full ACID and isInsertOnly(Table) to mean MM tables.
> This would accurately describe the semantics of the tables.

This message was sent by Atlassian JIRA

View raw message