[ https://issues.apache.org/jira/browse/FLINK-6225?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16006053#comment-16006053 ]
ASF GitHub Bot commented on FLINK-6225: --------------------------------------- Github user haohui commented on a diff in the pull request: https://github.com/apache/flink/pull/3748#discussion_r115927398 --- Diff: flink-connectors/flink-connector-cassandra/pom.xml --- @@ -176,5 +176,23 @@ under the License. </exclusion> </exclusions> </dependency> + <dependency> + <groupId>org.apache.flink</groupId> + <artifactId>flink-table_2.10</artifactId> + <version>${project.version}</version> + <scope>provided</scope> + </dependency> + <dependency> + <groupId>org.apache.flink</groupId> + <artifactId>flink-scala_2.10</artifactId> --- End diff -- You're correct. The problem is that in order to actually use the connector in a meaningful way users do need to depend on the packages, as `ExecutionEnvironment` and `StreamExecutionEnviornment` are exposed in the APIs: ``` $ javap flink-libraries/flink-table/target/classes/org/apache/flink/table/api/TableEnvironment.class ``` ``` public static org.apache.flink.table.api.scala.BatchTableEnvironment getTableEnvironment(org.apache.flink.api.scala.ExecutionEnvironment); public static org.apache.flink.table.api.java.BatchTableEnvironment getTableEnvironment(org.apache.flink.api.java.ExecutionEnvironment, org.apache.flink.table.api.TableConfig); public static org.apache.flink.table.api.java.BatchTableEnvironment getTableEnvironment(org.apache.flink.api.java.ExecutionEnvironment); ``` This is captured by the integration test in the patch. However, I do think these two dependency should be `test` instead of `provided`. At the very least the dependency does not get leaked to the users. @PangZhi can you please make the changes? > Support Row Stream for CassandraSink > ------------------------------------ > > Key: FLINK-6225 > URL: https://issues.apache.org/jira/browse/FLINK-6225 > Project: Flink > Issue Type: New Feature > Components: Cassandra Connector > Affects Versions: 1.3.0 > Reporter: Jing Fan > Assignee: Haohui Mai > Fix For: 1.3.0 > > > Currently in CassandraSink, specifying query is not supported for row-stream. > The solution should be similar to CassandraTupleSink. -- This message was sent by Atlassian JIRA (v6.3.15#6346)