[ https://issues.apache.org/jira/browse/FLINK-7016?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Robert Metzger updated FLINK-7016: ---------------------------------- Component/s: DataStream API DataSet API > Move inputFormat to InputFormatVertex from TaskConfig > ----------------------------------------------------- > > Key: FLINK-7016 > URL: https://issues.apache.org/jira/browse/FLINK-7016 > Project: Flink > Issue Type: Improvement > Components: DataSet API, DataStream API > Reporter: Xu Pingyong > Assignee: Xu Pingyong > > On batch case, InputFormat is put into TaskConfig, batch task gets it to read > data and job manager uses it to split splits from TaskConfig. > On streaming case, all configs are put into StreamConfig, but this > inputFormat is put into TaskConfig. > We can put InputFormat into InputFormatVertex, and batch task still gets > InputFormat from TaskConfig. It will be clear. -- This message was sent by Atlassian JIRA (v6.4.14#64029)