You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Apache Spark (JIRA)" <ji...@apache.org> on 2018/11/15 08:20:00 UTC
[jira] [Commented] (SPARK-26070) another implicit type coercion bug
[ https://issues.apache.org/jira/browse/SPARK-26070?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16687635#comment-16687635 ]
Apache Spark commented on SPARK-26070:
--------------------------------------
User 'uzadude' has created a pull request for this issue:
https://github.com/apache/spark/pull/23042
> another implicit type coercion bug
> ----------------------------------
>
> Key: SPARK-26070
> URL: https://issues.apache.org/jira/browse/SPARK-26070
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Affects Versions: 2.3.0, 2.4.0
> Reporter: Ohad Raviv
> Priority: Major
>
> looks like SPARK-22469 introduced a major bug to our system:
> {code}
> spark.sql("select '22222222222222222222' = 22222222222222222222BD").show()
> spark.sql("select '22222222222222222224' = 22222222222222222223BD").show()
> {code}
> which results:
> {noformat}
> +-----------------------------------------------------------------------------+
> |(CAST(22222222222222222222 AS DOUBLE) = CAST(22222222222222222222 AS DOUBLE))|
> +-----------------------------------------------------------------------------+
> | true|
> +-----------------------------------------------------------------------------+
> +-----------------------------------------------------------------------------+
> |(CAST(22222222222222222224 AS DOUBLE) = CAST(22222222222222222223 AS DOUBLE))|
> +-----------------------------------------------------------------------------+
> | true|
> +-----------------------------------------------------------------------------+
> {noformat}
> this causes downstream transformations to join together unrelated rows just because their ids are close.
> [~cloud_fan],[~liutang123] - could you please explain futher this remark:
> {noformat}
> // There is no proper decimal type we can pick,
> // using double type is the best we can do.
> // See SPARK-22469 for details.
> {noformat}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org