[ https://issues.apache.org/jira/browse/SPARK-30893?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17041617#comment-17041617 ]
L. C. Hsieh commented on SPARK-30893: ------------------------------------- hmm, an action like `df.collect` should re-evaluate if it is not cached, right? If so, changing a config causing the results different sounds reasonable too? Although I agree that it is error-prone in practice. We should avoid inconsistency among different SQL configs, for example some configs can change behaviors across `df.collect` calls but some are not. If there is a consistent rule about whether SQL configs can change behaviors across different calls of same action, it is better. > Expressions should not change its data type/behavior after it's created > ----------------------------------------------------------------------- > > Key: SPARK-30893 > URL: https://issues.apache.org/jira/browse/SPARK-30893 > Project: Spark > Issue Type: Bug > Components: SQL > Affects Versions: 3.0.0 > Reporter: Wenchen Fan > Priority: Critical > > This is a problem because the configuration can change between different > phases of planning, and this can silently break a query plan which can lead > to crashes or data corruption. -- This message was sent by Atlassian Jira (v8.3.4#803005) --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org