[ https://issues.apache.org/jira/browse/FLINK-1819?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14659466#comment-14659466 ]
ASF GitHub Bot commented on FLINK-1819: --------------------------------------- Github user sachingoel0101 commented on the pull request: https://github.com/apache/flink/pull/966#issuecomment-128230692 I'm inclined towards the Abstract class too. If we're saying something is context aware, we better provide access to it IMO. It doesn't make sense to ask the user to write it themselves and guarantee that we'd just call the relevant methods at runtime. As for the names, wouldn't it be better to maintain consistency in naming? All the * rich * functions are named `RichFunctions`, which is why I decided to name these `RichFormat`s too. I do agree though. The IO formats are already pretty * rich * in the sense that they allow the user to control everything that happens at runtime. > Allow access to RuntimeContext from Input and OutputFormats > ----------------------------------------------------------- > > Key: FLINK-1819 > URL: https://issues.apache.org/jira/browse/FLINK-1819 > Project: Flink > Issue Type: Improvement > Components: Local Runtime > Affects Versions: 0.9, 0.8.1 > Reporter: Fabian Hueske > Assignee: Sachin Goel > Priority: Minor > Fix For: 0.9 > > > User function that extend a RichFunction can access a {{RuntimeContext}} > which gives the parallel id of the task and access to Accumulators and > BroadcastVariables. > Right now, Input and OutputFormats cannot access their {{RuntimeContext}}. -- This message was sent by Atlassian JIRA (v6.3.4#6332)