[jira] [Updated] (HIVE-12631) LLAP: support ORC ACID tables

["Sergey Shelukhin (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HIVE-12631?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Sergey Shelukhin updated HIVE-12631:
------------------------------------
    Attachment: HIVE-12631.27.patch

Updated the patch. Not sure why the config was set in UpdateDeleteSemanticAnalyzer so I commented that out for now. I looked a bit at the CVB-VRB-CVB-VRB conversion, given that handling a selected vector after ACID reader requires copying stuff, it doesn't seem ideal. Can be handled in a followup. Either a selected vector can be added to CVB and the ACID merger thing made operate on both (the code is common between the two), or LLAPRecordReader can be enabled to accept VRBs directly.

> LLAP: support ORC ACID tables
> -----------------------------
>
>                 Key: HIVE-12631
>                 URL: https://issues.apache.org/jira/browse/HIVE-12631
>             Project: Hive
>          Issue Type: Bug
>          Components: llap, Transactions
>            Reporter: Sergey Shelukhin
>            Assignee: Teddy Choi
>         Attachments: HIVE-12631.10.patch, HIVE-12631.10.patch, HIVE-12631.11.patch, HIVE-12631.11.patch, HIVE-12631.12.patch, HIVE-12631.13.patch, HIVE-12631.15.patch, HIVE-12631.16.patch, HIVE-12631.17.patch, HIVE-12631.18.patch, HIVE-12631.19.patch, HIVE-12631.1.patch, HIVE-12631.20.patch, HIVE-12631.21.patch, HIVE-12631.22.patch, HIVE-12631.23.patch, HIVE-12631.24.patch, HIVE-12631.25.patch, HIVE-12631.26.patch, HIVE-12631.27.patch, HIVE-12631.2.patch, HIVE-12631.3.patch, HIVE-12631.4.patch, HIVE-12631.5.patch, HIVE-12631.6.patch, HIVE-12631.7.patch, HIVE-12631.8.patch, HIVE-12631.8.patch, HIVE-12631.9.patch
>
>
> LLAP uses a completely separate read path in ORC to allow for caching and parallelization of reads and processing. This path does not support ACID. As far as I remember ACID logic is embedded inside ORC format; we need to refactor it to be on top of some interface, if practical; or just port it to LLAP read path.
> Another consideration is how the logic will work with cache. The cache is currently low-level (CB-level in ORC), so we could just use it to read bases and deltas (deltas should be cached with higher priority) and merge as usual. We could also cache merged representation in future.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)