You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Ivan Bessonov (Jira)" <ji...@apache.org> on 2022/12/15 11:35:00 UTC
[jira] [Commented] (IGNITE-17467) Find out Row id is not possible for MvPartitionStorage
[ https://issues.apache.org/jira/browse/IGNITE-17467?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17648006#comment-17648006 ]
Ivan Bessonov commented on IGNITE-17467:
----------------------------------------
RowId values are required to re-acquire exclusive locks or transactions state after node restart or leader change, basing them on write intents.
> Find out Row id is not possible for MvPartitionStorage
> ------------------------------------------------------
>
> Key: IGNITE-17467
> URL: https://issues.apache.org/jira/browse/IGNITE-17467
> Project: Ignite
> Issue Type: Bug
> Reporter: Vladislav Pyatkov
> Priority: Major
> Labels: ignite-3
>
> Row id is returned only for insert operation, but if you want to find a row id by Binary row is not possible.
> I sure, it has to be possible through scan operation. In that case, the scan operation can be used as a scan index.
> I would like to see API like:
> {code:java}
> Cursor<Tuple<RowId, BinaryRow>> scan(Predicate<BinaryRow> keyFilter, UUID txId) throws TxIdMismatchException, StorageException;
> Cursor<Tuple<RowId, BinaryRow>> scan(Predicate<BinaryRow> keyFilter, Timestamp timestamp) throws StorageException;
> {code}
>
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)