You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Dongjoon Hyun (Jira)" <ji...@apache.org> on 2020/03/16 22:52:08 UTC

[jira] [Updated] (SPARK-26001) Reduce memory copy when writing decimal

     [ https://issues.apache.org/jira/browse/SPARK-26001?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Dongjoon Hyun updated SPARK-26001:
----------------------------------
    Affects Version/s:     (was: 3.0.0)
                       3.1.0

> Reduce memory copy when writing decimal
> ---------------------------------------
>
>                 Key: SPARK-26001
>                 URL: https://issues.apache.org/jira/browse/SPARK-26001
>             Project: Spark
>          Issue Type: Improvement
>          Components: SQL
>    Affects Versions: 3.1.0
>            Reporter: caoxuewen
>            Priority: Major
>
> this PR fix 2 here:
> - when writing non-null decimals, we not zero-out all the 16 allocated bytes. if the number of bytes needed for a decimal is greater than 8. then we not need zero-out between 0-byte and 8-byte. The first 8-byte will be covered when writing decimal.
> - when writing null decimals, we not zero-out all the 16 allocated bytes. BitSetMethods.set the label for null and the length of decimal to 0. when we get the decimal, will not access the 16 byte memory value, so this is safe.



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

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org