[
https://issues.apache.org/jira/browse/HIVE-23067?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Marton Bod updated HIVE-23067:
------------------------------
Description:
To reduce the number of database calls and network roundtrips, we could use
more batching in TxnHandler, where currently in many places we call insert
commands in loops sequentially.
Some examples:
* openTxns (TXNS, REPL_TXN_MAP)
* commitTxn (COMPLETED_TXN_COMPONENTS)
* replTableWriteIdState (TXN_TO_WRITE_ID)
* allocateTableWriteIds (TXN_TO_WRITE_ID)
* ....
was:
To reduce the number of database calls and network roundtrips, we could use
more batching in TxnHandler, where currently in many places we call insert
commands in loops sequentially.
Some examples:
* openTxns (TXNS, REPL_TXN_MAP
* commitTxn (COMPLETED_TXN_COMPONENTS)
* replTableWriteIdState (TXN_TO_WRITE_ID)
* allocateTableWriteIds (TXN_TO_WRITE_ID)
* ....
> Use batch inserts in TxnHandler
> -------------------------------
>
> Key: HIVE-23067
> URL: https://issues.apache.org/jira/browse/HIVE-23067
> Project: Hive
> Issue Type: Improvement
> Reporter: Marton Bod
> Assignee: Marton Bod
> Priority: Major
>
> To reduce the number of database calls and network roundtrips, we could use
> more batching in TxnHandler, where currently in many places we call insert
> commands in loops sequentially.
> Some examples:
> * openTxns (TXNS, REPL_TXN_MAP)
> * commitTxn (COMPLETED_TXN_COMPONENTS)
> * replTableWriteIdState (TXN_TO_WRITE_ID)
> * allocateTableWriteIds (TXN_TO_WRITE_ID)
> * ....
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
