[ https://issues.apache.org/jira/browse/FLINK-11067?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16719091#comment-16719091 ]
Dawid Wysakowicz edited comment on FLINK-11067 at 12/12/18 3:29 PM: -------------------------------------------------------------------- I think as user should only import either {{flink-tabe-api-java}} or {{flink-tabe-api-scala}} never both of them we can have: {code} // in flink-table-api-java module package org.apache.flink.table.api; public final class TableEnvironment { static org.apache.flink.table.api.java.BatchTableEnvironment getTableEnvironment(BatchEnvironment); static org.apache.flink.table.api.java.StreamTableEnvironment getTableEnvironment(StreamEnvironment); private TableEnvironment() { } } // in flink-table-api-scala module package org.apache.flink.table.api; object TableEnvironment { def getTableEnvironment(BatchEnvironment): org.apache.flink.table.api.scala.BatchTableEnvironment = ... def getTableEnvironment(StreamEnvironment): org.apache.flink.table.api.scala.StreamTableEnvironment = .... } {code} The rest would remain the same as in [~twalthr] suggestion. I think that would retain backwards compatibility. was (Author: dawidwys): I think as user should only import either {{flink-tabe-api-java}} or {{flink-tabe-api-scala}} never both of them. I think we can have: {code} // in flink-table-api-java module package org.apache.flink.table.api; public final class TableEnvironment { static org.apache.flink.table.api.java.BatchTableEnvironment getTableEnvironment(BatchEnvironment); static org.apache.flink.table.api.java.StreamTableEnvironment getTableEnvironment(StreamEnvironment); private TableEnvironment() { } } // in flink-table-api-scala module package org.apache.flink.table.api; object TableEnvironment { def getTableEnvironment(BatchEnvironment): org.apache.flink.table.api.scala.BatchTableEnvironment = ... def getTableEnvironment(StreamEnvironment): org.apache.flink.table.api.scala.StreamTableEnvironment = .... } {code} The rest would remain the same as in [~twalthr] suggestion. I think that would retain backwards compatibility. > Port TableEnvironments to Java > ------------------------------ > > Key: FLINK-11067 > URL: https://issues.apache.org/jira/browse/FLINK-11067 > Project: Flink > Issue Type: Sub-task > Components: Table API & SQL > Reporter: Timo Walther > Assignee: Dawid Wysakowicz > Priority: Major > > This task includes porting {{TableEnvironment}}, {{StreamTableEnvironment}}, > {{BatchTableEnvironment}} to Java. API-breaking changes need to be avoided > and discussed. Some refactoring and clean up might be necessary. -- This message was sent by Atlassian JIRA (v7.6.3#76005)